Identification of chloroplast envelope proteins in close physical proximity to a partially translocated chimeric precursor protein.

Translocation intermediates of the chimeric protein precursor Oee1-Dhfr were generated and used to identify envelope components in close proximity to the arrested precursor. The translocation of Oee1-Dhfr across the chloroplast envelope can be arrested at low ATP levels or by prebinding the fusion precursor with anti-Dhfr IgGs. The arrested Oee1-Dhfr precursor appears to span both the outer and inner envelope membranes. Translocational arrest of Oee1-Dhfr by low ATP levels was reversible, and import was restored upon resupplementation with higher ATP levels. Chemical cross-linking and co-immunoprecipitation with monospecific antibodies indicate that two outer envelope membrane proteins (Com44 and Com70) and at least one inner envelope protein (Cim44 and Cim97) were found to be in close proximity to Oee1-Dhfr during translocation. The Com70 protein was further studied and additional evidence for its role in chloroplast protein import is presented.

Two other strategies, however, were successful in arresting import. Waegemann and Soll (1991) identified translocation intermediates of the precursor of the small subunit of ribulose-1,5-bisphosphate carboxylase using isolated outer envelope vesicles. Schnell and Blobel (1993) identified two import intermediates of an engineered chloroplast protein precursor, one was specifically bound to the envelope, and the other spanned across both envelope membranes. The elusive nature of chloroplastic translocation intermediates suggests that the plastid protein import mechanism may possess some characteristics unique from other transport systems.
Therefore, effective means for the production of import intermediates should be explored extensively in the chloroplast.
In this paper, the translocation of a chimeric precursor protein, Oeel-Dhfr,' was arrested across both envelope membranes by using relatively low ATP levels or by preincubating the precursor with anti-Dhfr IgGs. Upon establishing the nature of the arrested Oeel-Dhfr, we proceeded to use them to identify envelope proteins that were in close proximity to the translocating precursor. One of these components, Com70, is a heat shock-related protein (formerly designated SCE70 in KO et al. (1992) and Wu and KO (1993)) and the present study provides evidence that it probably plays a role in protein import.

MATERIALS AND METHODS
Production ofAntibodies-Anti-Com70 and anti-Dhfr IgGs were prepared as reported ( KO et al., 1992;Wu and KO, 1993). Specific IgGs against the COOH terminus of the chloroplast envelope proteins (the 37-kDa Cim37, the two 44-kDa Com44Kim44, and the 97-kDa Cim97) were similarly generated. The plasmid vector pGEMEX-1 (Promega) and Escherichia coli strain JMlOg(DE3) were used to facilitate overexpression. Protein inclusion bodies were purified according to a procedure in Sambrook et al. (1989). Denatured fusion proteins were prepared by SDS-PAGE. The generation and purification of rabbit IgGs was according to Chua et al. (1982). Control IgGs were purified from preimmune sera collected prior to immunization.
Preparation of Radiolabeled Precursors-The DNA templates encoding the Oeel (from Arabidopsis thaliana) and Oeel-Dhfr precursor proteins were cloned into pGEM4 (Promega) (for details, see KO and Cashmore (1989)). Radiolabeled proteins were synthesized using in ~~ ~ The abbreviations used are: Oeel-Dhfr, the fusion protein between the transit peptide of the 33-kDa oxygen-evolving protein and mouse dihydrofolate reductase; PAGE, polyacrylamide gel electrophoresis; Cim, chloroplast inner envelope membrane protein; Com, chloroplast outer envelope membrane protein; DSP, dithiobis (succinimidylpropionate); DTSSP, 3,3'-dithiobis (sulfonylsuccinimidyl-propionate); I&, immunoglobulin G; PMSF, phenylmethylsulfonyl fluoride; RBCS, precursor of the small subunit of ribulose-l,5-bisphosphate carboxylase; SDS, sodium dodecyl sulfate. vitro transcription and translation systems described by Melton et al. (1984) and Erickson and Blobel (1983). The templates were transcribed using SP6 RNA polymerase (Promega, Pharmacia Biotech Inc., Life Technologies, Inc.). Import into Chloroplasts and Subfractionation Procedures-Intact pea chloroplasts were isolated from 11-14-day-old seedlings as described by Bartlett et a1.(1982) or Cline et al. (1985). The growth conditions were identical to those described earlier (KO and Cashmore, 1989). The in vitro import assays were assembled and carried out according to Bartlett et al. (1982). Binding assays were performed using the ionophore nigericin (Calbiochem) (Cline et al., 1985). Reisolated intact chloroplasts were given further treatments according to Smeekens et al. (1986). Subfractionation of chloroplasts was carried out according to Keegstra and Yousif (1986). Various conditions were modified and will be indicated accordingly.
Chemical Cross-linking and Immunoprecipitation-The cross-linking experiments were performed according to Sanders et al. (1992). The cross-linked products were immunoprecipitated as described by Scherer et al. (1990).
Preparation of Total Cell Membranes from Various Plant Tissues-Tissues were collected from 6-week-old pea plants (Pisum satiuun) and 10-week-old Brassica napus (cv. Topas) plants. Young pea leaves were selected from the shoot and mature leaves from further down the stem. Seeds, flowers, and roots were also harvested from pea seedlings. Etiolated pea cotyledons were collected from 10-day-old dark-grown seedlings. In B. napus, young leaves of a maximum size of 2 cm were selected from the shoot and mature leaves (minimum size of 7 cm) from further down the stem. Flowers, roots, and 21 day-old seeds were also harvested. Mature seeds were taken from the B. napus plants as the siliques started to turn yellow and began drying out. All tissues (B. napus and I? sativum) were immediately frozen in liquid nitrogen and stored at -80 "C until use. Total cell membranes were prepared according to Sutton et al. (1987).
Electrophoresis, Protein Blotting, and Fluorography-Samples from in vitro wheat germ translations, various binding and import reactions, immunoprecipitated complexes, and total cell membrane samples were analyzed by SDS-PAGE (Laemmli, 1970). Proteins resolved by SDS-PAGE were transferred, probed, and processed as described by Towbin et al. (1979) and Hoffman et al. (1987). The avidin-biotin detection system (Vector Laboratories, Burlingame, CA) or anti-rabbit IgG alkaline phosphatase conjugates (Sigma) were used. Selected gels were prepared for fluorography using ENHANCE (DuPont NEN) and exposed to Kodak XAR x-ray films. The LKB Ultroscan XL laser densitometer was used to quantitate and normalize the resulting Western blots and the fluorographs of various import experiments.

I)-anslocation
Intermediates of Oeel-Dhfr-Two strategies were used to arrest the import of Oeel-Dhfr across the chloroplast envelope, prebinding the precursor with anti-Dhfr IgGs and low ATP levels. Oeel-Dhfr (molecular mass of 30 kDa) is processed to distinct stromal intermediate (25 kDa) and thylakoid mature forms (28 kDa) in a typical import assay (1 mM ATP) (KO and Cashmore, 1989). Oeel-Dhfr was selected for this study because it possesses two useful features. One feature is the lengthy transit peptide (85 residues), which is sufficiently long to span both envelope membranes and allows us to follow the progress of import. The other feature is its slower import rate relative to authentic Oeel or other precursors.
Prebinding precursors with anti-Dhfr IgGs occurred prior to the addition of chloroplasts. Bound IgGs inhibited Oeel-Dhfr import (Fig. 1B), whereas preimmune IgGs did not significantly alter import (Fig. l A ) even though it did affect binding at higher concentrations. Binding decreased substantially while the stromal intermediate and mature forms remained largely unaffected (Fig. lA, lanes 3 4 ) . Inhibition of binding a t high control IgG levels may be attributed to nonspecific effects. The extent of import inhibition by anti-Dhfr IgGs was dependent on the prebinding IgG levels used (Fig. 1 B , lanes 2 4 ) . Anti-Dhfr IgGs were specific for Dhfr with no cross-reactivity to any chloroplast proteins (Wu and KO, 1993) and did not significantly affect the import of authentic Oeel precursors (Fig. 1C). Multiple processed Oeel-Dhfr forms were detected in reactions us-Import Apparatus Only the precursor form was detected with anti-Dhfr IgG prebinding and nigericin (Fig. v), lanes 2 4 ) . Therefore, the IgG preparations did not contribute to the observed cleavage of Oeel-Dhfr. The multiple processing events appear to take place in the chloroplast, suggesting that the precursor has crossed the envelope and halted at various points along its import pathway. The main processed form has a molecular mass of 28 kDa (designated Oeel-Dhfr28) and is clearly different from the stromal intermediate. The time course study showed that in the absence of IgG prebinding, Oeel-Dhfr was imported at a slower rate than Oeel (Fig. 2B) and advanced only into the stroma at 2 min ( Fig. 2 B , lane 2 ) . The mature protein appeared at 5 min ( Fig. 2 B , lane 3). Oeel was imported efficiently with significant amounts of the mature form appearing at 2 min, and intermediate forms were not observed ( Fig. 2 A , lanes 1-81. In contrast, IgG prebinding abolished both the stromal intermediate and the mature form of Oeel-Dhfr (Fig. 2C, IgG concentration equivalent to lane 3 in Fig. 1B). Only the precursor and Oeel-Dhfr28 were present. Oeel-Dhfr28 appeared a t 5 min and accumulated to a significant level by 30 min (Fig. 2C).
Reisolated intact chloroplasts were fractionated into outer Protein Import Apparatus and inner envelopes, stroma and thylakoids to localize Oeel-Dhfr28 (Fig. 3B). The typical fractionation pattern was observed for Oeel-Dhfr in the absence of IgGs (Fig. 3A, lanes 2-5). In the presence of IgG prebinding, the precursor was found associated with both envelope fractions but mostly with the inner membranes ( Fig. 3 A , lanes 6-7). Other forms were not detected in the stroma or thylakoid (Fig. 3 A , lanes 8-9).
The configuration of Oeel-Dhfr and Oeel-Dhfr28 in the envelope was examined using thermolysin and trypsin treatments. Protein moieties exposed on the outside surface are susceptible to both proteases. Since trypsin can penetrate the outer membrane, polypeptides exterior to the inner membrane will also be trypsin sensitive (Cline et al., 1984;Joyard et al., 1983). Only the precursor was susceptible to both proteases in the control assays ( Fig. 3B, lanes 3 and 5). The stromal intermediate and the mature form were both resistant and are thus located internal to the inner envelope. Thermolysin treatment of chloroplasts from import assays conducted with IgG prebinding showed that a large part of the precursor and Oeel-Dhfr28 was protected from degradation. Three distinct degradation products with molecular mass values of 28, 26, and 23.5 kDa were observed (Fig. 3B, lane 4 1. The 26-kDa product is slightly higher than the stromal intermediate, and the 23.5-kDa form is slightly larger than the mature form (Fig. 3B, lane 4 ) . Trypsin treatment resulted in two distinct degradation products of approximately 24 and 23 kDa. The 24 kDa band is slightly higher than the mature form, and the other is about the same size as the mature form (Fig. 3B, lane 6). These results suggest that translocation of Oeel-Dhfr was halted at various points along its import route and that a portion of the precursors spans both envelope membranes becoming accessible to internal processing events.
Oeel-Dhfr translocation can also be halted by low exogenous ATP levels. This series of assays was carried out under dim green light, a condition that prevents light-dependent chloroplastic ATP production. The maximal ATP contribution from 1 1 2 3 4 5 8 1 FIG. 3. Localization of Oeel-Dhfr28 in the chloroplast envelope.A, subfractionation of the chloroplast. Lane 1 represents a control import assay with the typical Oeel-Dhfr import pattern.
Lanes 2-5 represent the outer membrane, the inner membrane, the stroma, and the thylakoid fraction of the control assay, respectively. Fractions in lanes 6-9 are in the same order as in lunes 2-5, except that the fractions were derived from assays where Oeel-Dhfr was preincubated with 200 pg of anti-Dhfr IgGs as in Fig. 2B. B , postuptake treatment with proteases. Lunes 1-2, untreated; lunes 3 4 , thermolysin treatment; lunes 5-6, trypsin treatment. Assays in lunes 1 , 3 , and 5 were carried out in the absence of antibody prebinding. Reactions in lunes 2, 4, and 6 represent the assays with anti-Dhfr IgG prebinding. Import conditions were as in Fig. 1 the wheat germ and the in vitro translation reactions was estimated to be 25 PM. The typical pattern of Oeel-Dhfr import was observed with ATP levels 2 250 PM. Oeel-Dhfr was only partially imported with < 250 PM ATP. A distinct form of Oeel-Dhfr (designated Oeel-Dhfr26) with a molecular mass of 26 kDa was observed (Fig. 4 B , top row, lanes [2][3][4][5] and was most apparent at 50-75 PM ATP (lanes 2-3) but disappeared at ATP levels > 250 PM (lanes . Decreases in the amount of Oeel-Dhfr26 were accompanied by a concomittant increase in the level of the imported forms. Both the precursor and Oeel-Dhfr26 co-fractionated exclusively with the envelope and were not detected in the other fractions (Fig. 4C, lanes 4-6). At 25 PM ATP, Oeel-Dhfr appeared only as a full precursor (Fig. 4B, top row, lane 1 ) and was the only form detected in the envelopes (Fig. 4C, lanes 1 3 ) . The usual fractionation pattern was observed with 1 mM ATP assays (Fig. 4C, lanes 7-9). Thermolysinsensitive Oeel import intermediates, although in much lower amounts, were also observed in a parallel experiment (Fig. 44, lanes 2 4 of each row). The typical Oeel import pattern was observed at 25 PM (Fig. 44, lane 1 ) or at 2250 PM ATP (Fig. 44,  lanes 5-8). Therefore, Oeel-Dhfr26 is located in the envelope consistent with our expectations of an intermediate.
Translocational arrest of Oeel-Dhfr a t 50 PM ATP appears to be reversible by higher levels of ATP. Resupplementation with 25 PM ATP did not yield any detectable changes relative to the control (Fig. 4 0 , lane 2 uersus lane 1 ). Import was restored only when ATP levels 2100 PM were resupplied (Fig. 40, lanes 3 4 ). Oeel-Dhfr26 disappeared with a concomittant appearance of the usual imported forms suggesting that the arrested precursors can be chased into the chloroplast by higher levels of ATP thus representing productive intermediates.
Oeel-Dhfr26 appears to be processed a t a site different from the stromal intermediate. Moreover, Oeel-Dhfr26 and the precursor were both thermolysin-sensitive (Fig. 4 B , bottom row,  lanes 2-5), indicating that a large part of the arrested precursor or Oeel-Dhfr26 was exposed on the exterior; therefore, the precursors were probably arrested early in its passage across the envelope. The precursor may have traversed the envelope but was not fully accessible to stromal processing events. The extent of translocation was determined by analysis of the small pro- teolytic products. Chloroplasts were first depleted of ATP by a dark incubation and then supplemented with 4 1.1~ to 1 mM ATP. The typical Oeel-Dhfr import pattern was observed at 1 mMATP (Fig. 5, A-C, lune 9 ). The full precursor was the main form detected a t ATP levels 5100 PM (Fig. 5A, lunes 1-8). Oeel-Dhfr26 appeared as a distinct band at 60-100 PM ATP. (Fig. 5A, lanes  7-8). Lower ATP levels (100-10 PM) gave rise to distinct 8-10-kDa protease-resistant products (Fig. 5, B-C, lunes 4-81, suggesting that at least a segment of the precursor, presumably the NH, terminus, has crossed the envelope and into the stroma. Two distinct protease-resistant products were detected (Fig.  5D). One product may have been derived from partially trans- reactions were posttreated with thermolysin, and C represents the results of posttreatment with trypsin. Lanes 1-9 represent assays with 4, 6,8,10,20,35,60, 100, and 1000 ~M A T P , respectively. Total chloroplast protein profiles were analyzed.p, i, i*, and m are as in Fig. 4. D , analysis of thermolysin-and trypsin-generated proteolytic products using a 20% polyacrylamide gel. Samples from lunes 7 and 8 ofpunels B and C were analyzed, and the resulting fluorogram is presented in lunes 1-2 and 3 4 of panel D, respectively. The arrows highlight the position of the proteolytic products. located full precursors and the other from Oeel-Dhfr26. However, at present we cannot determine the direct relationship between the nondegraded and degraded forms. Decreases in ATP levels (<35 p~) resulted in substantial decreases in the amount of small proteolytic products with a concomittant increase in the amount of a thermolysin-resistant full precursor (Fig. 5B, lunes   1-6). This precursor form, however, was susceptible to trypsin ( Fig. 5C, lunes 1-61, suggesting that ATP levels 510 w may support translocation across the outer envelope only presumably into the intermembrane space and that ATP levels >10 p~ support partial translocation across both membranes.

Envelope Proteins in Close Proximity to Ranslocating Oeel-Dhfr Precursors-
The low ATP-generated Oeel-Dhfr translocation intermediates are presumably associated with the import machinery and can thus be used to identify components in close Antibody specificity was monitored by Western blotting. Lanes 1-5 in B-D represent the total envelope, the outer envelope, the inner envelope, the stroma, and the thylakoidal fractions, respectively. Total immunoreactive protein profiles of the outer and the inner envelope (A, lanes I and 2, respectively) were detected with IgGs against total pea chloroplast envelope proteins. Mr, relative molecular mass markers in kilodaltons.
The cross-linking experiments suggest that sufficient amounts of precursors were arrested in the envelope at low ATP levels allowing efficient cross-linking to envelope components of the import machinery, whereas cross-linked complexes were not precipitated by any of the IgGs when the ATP level was elevated to 1 mM (Fig. 7B, lanes 1 and 3-5). DSP was used in the 1 mM ATP assays because Oeel-Dhfr can be cross-linked to Com70, Com44/Cim44, and Cim97 with DSP, whereas DTSSP did not form cross-links with every component. As expected, precursors were not immunoprecipitated by preimmune or anti-Cim37 IgGs (Fig. 7B, lanes 2 and 6 ) . The results demonstrate that associatioddisassociation of Oeel-Dhfr from Com70, Com44Kim44, and Cim97 is ATP-dependent. Low ATP levels may stabilize the association, whereas higher ATP levels promote dissociation. This hypothesis was further confirmed by experiments with the authentic Oeel precursor. The Oeel precursor was imported more efficiently even at 100 p~ ATP. Therefore, the interaction between the Oeel precursor and the protein import machinery is probably short lived even at low ATP levels. The results demonstrate that under these conditions Oeel was not cross-linked with any of the envelope proteins tested. For example, no Oeel in any form was precipitated by anti-Com70 IgGs despite the use of cross-linkers (Fig. 7C,  lanes 3-4). Impairment of Protein Import by Anti-Corn70 IgGs-Of the envelope proteins identified above, Corn70 is the most accessible to immediate additional characterization due to its external location. Com7O's accessibility can be seen in experiments displaying a correlation between the amount of Corn70 removed by thermolysin and the level of activity at the early stages of protein translocation, but a direct relationship cannot be defined since thermolysin is not Com70-specific and can therefore affect other protease-sensitive components.2 A more specific approach is to measure the effects of anti-Com70 IgGs on the import process. Chloroplasts were preincubated with IgGs on ice for 45 min prior to the addition of precursors. Authentic Oeel precursors were chosen because of their rapid advancement from the binding step to the processed mature form such that any slight effect caused by IgGs can be detected and quantitated more readily. Preimmune IgGs did not have a significant effect on Oeel import (Fig. 8A). Increasing preimmune IgG concentration from 1 to 2 pg/pg of chlorophyll reduced the level of Oeel import by approximately 10% relative to the mock control (Fig. 8A, lane 3, and 8C). The same concentration of anti-Com70 IgGs (2 pg/pg chlorophyll) resulted in a decrease of approximately 60% (Fig. 8B, lane 3;  is probably attributed to the interfering effect of the IgG molecule. Corn70 in Various Plant Tissues-Western blot analysis of Corn70 in various types of plant tissue and developmental stages was conducted to determine if the distribution pattern correlated with the plastid's protein import activity (Fig. 9, A and B ) . The following trends were observed for the two plant species tested; seeds and flowers contain the highest levels, roots exhibit intermediate amounts, and leaves contain the lowest amounts. Within each type of tissue, the level of Corn70 was higher in younger than older stages, e.g. etiolated cotyledons or young leaves uersus old leaves, or 21-day-old seeds versus mature seeds. The relative amounts of Corn70 were normalized against total envelope protein content. In contrast to Com70, Cim37 levels remained relatively ~n c h a n g e d .~ Therefore changes in the amount of Com7O were not due to changing amounts of envelope membranes present in the samples. These results show that the expression of Corn70 is perhaps developmentally regulated and is in close correlation with the predicted import activity of the tissue.

DISCUSSION
Import intermediates are valuable tools for studying the passage of a protein across the envelope membrane and for identifying putative components involved. The first part of this study is focused on the generation of translocation intermediates of an engineered chloroplast precursor protein. Upon establishing the nature and configuration of the arrested precursors, we in turn proceeded to use them to identify envelope proteins in close physical proximity to the intermediates. This  1-6) and B. napus tissues (lanes 7-12). B, Corn70 content relative to the total plastid envelope proteins (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12) represent the normalized results of lane 1-12 in A, respectively). Total cell membranes were from young (lanes 1 and 7 ) and mature leaves (lunes 2 and 8), flowers (lanes 3 and 91, roots (lanes 4 and lo), mature seeds (lanes 5 and 12), 10-day dark-grown cotyledons (lane 6) and 21-day-old seeds (lane 11 ). The blots were analyzed by densitometry, and relative amounts of Corn70 were normalized against total envelope membrane proteins. work culminated in the development of a model for the translocation of Oeel-Dhfr across the chloroplast envelope and the putative components involved (Fig. 10).
Two different strategies, prebinding the precursor with an IgG molecule and low ATP concentrations, were used to arrest Oeel-Dhfr import. Both approaches led to precursors that span both outer and inner envelope membranes. Prebinding Oeel-Dhfr with IgGs inhibited transport as is evident by the disappearance of the imported forms (Fig. lB) and the appearance of a form that fractionated with the inner envelope (Fig. 3A). Arrested precursors were processed a t a number of cryptic sites giving rise to different sized species with the main one being a 28-kDa form (Oeel-Dhfr28). Since purified IgGs showed no protease activity toward Oeel-Dhfr, aberrant processing most likely occurred inside and was dependent on import activity. Furthermore, multiple processed forms were not detected with nigericin. This suggests that at least the transit peptide was exposed to locations internal to the outer envelope. The proteolysis results further suggest that a portion of the arrested precursors span across both the outer and inner envelopes at various points in the import pathway (Fig. lOA, molecule 3).
Low ATP-generated translocation intermediates span both envelopes in a similar manner except that the stage of arrest is earlier and reversible in nature. The 26-kDa intermediate (Oeel-Dhfr26) (Fig. 4B) appeared at ATP levels <250 p~. Oeel-Dhfr26 is clearly different from the stromal intermediate for the following reasons, 1) Oeel-Dhfr26 appears to be processed a t a different site resulting in a molecular mass slightly higher than the stromal intermediate, 2) Oeel-Dhfr26 is thermolysinsensitive, 3) Oeel-Dhfr26 disappeared at high levels of ATP while the stromal intermediate accumulated (Fig. 4B), and 4) Oeel-Dhfr26 co-fractionated with the envelope rather than the stroma (Fig. 4 0 . Although the arrested precursors remained largely proteasesensitive, 8-10-kDa segments have crossed the envelope and became protease-protected (Fig. 5). Unlike the IgG-generated intermediates, these low ATP-arrested precursors remained largely at the surface, and a smaller segment, presumably the NH,-segment, had entered the stroma (Fig. 10, molecule 1).
Translocational arrest using low ATP conditions appears to be productive and reversible based on a number of observations. First, the generation of translocation intermediates is dependent on the low ATP levels and disappeared a t higher ATP concentration. Second, the 8-10-kDa protease-protected products were present only a t low ATP ranges and were not detected in abundance at higher ATP levels or when the ATP level was 510 p~. Interestingly, a form of the full precursor was resistant to thermolysin but was susceptible to trypsin a t 10 p~ ATP (Fig. 5, B versus C , lanes 6-8), which suggests that it has crossed the outer membrane toward the intermembrane space (Fig. 10, molecule 2). Finally, the ability to support subsequent import of the intermediates by ATP re-supplementation suggests that a sufficient quantity of the arrested precursors are retained along the import pathway.
It is necessary to establish that the precursor spans both envelopes and that arrest is reversible to ensure that it will be in close proximity to as many envelope components as possible. The simultaneous presence of more than one type of intermediate also enhances the probability of chemical cross-linking and subsequently identify the envelope components with specific IgGs. This approach has helped us identify a t least three such candidates, Com70, Com44/Cim44m, and Cim97 (Fig.  1OB). The cross-linkinghmmunoprecipitation reactions are specific for the following reasons; 1) the IgGs are specific for their respective envelope proteins. These IgGs do not cross-react with Oeel-Dhfr nor with any other chloroplast proteins; 2) anti-Cim37 IgGs did not precipitate Oeel-Dhfr, indicating that cross-linking resulted in specific protein complexes rather than the formation of nonspecific networks of cross-linked proteins or membrane vesicles that contained everything. Cim37 is appropriate for this purpose due to its relative abundance in the envelope; 3) cross-linked complexes were not immunoprecipitated when import was enhanced by 1 mM ATP.
Even though these components formed cross-links with the arrested precursors, there were some interesting differences between DTSSP, the membrane-impermeable cross-linker, and DSP, a membrane-permeable cross-linker. For instance, Corn70 can be cross-linked to the precursor with DTSSP, suggesting that Corn70 may function externally at an early stage of the import process, whereas DSP cross-linked Corn70 to both precursor and Oeel-Dhfr26. Although the exact reason for this difference is presently not known, it is possible that Oeel-Dhfr26 was engaged in such a way that cross-linking was achieved with DSP only. Another interesting result is that significantly lower amounts of Oeel-Dhfr were cross-linked to Com44/Cim44 with DTSSP rather than DSP, a result that correlates with the amount of Corn44 present in the outer membrane and the amount of Cim44 in the inner membrane.4 Both Corn44 and Cim44 appear to be located at the import pathway and can be cross-linked to the intermediates. However, we do not yet have direct evidence differentiating between the possibilities of both being involved or just one, therefore we decided to refer to them as Com44/Cim44. Among the three envelope proteins identified, Com70, a hsp70-related component, is located in an outer envelope com-K. KO, D. Budd, C. W. Wu, F. Siebert, L. Kourtz, and Z. W. KO, submitted for publication.