Proc. Nati. Acad. Sci. USAVol. 86, pp. 886-890, February 1989Botany

Effect of mutations on the binding and translocation functions of achloroplast transit peptide

(nigericin/ribulose 1,5-bisphosphate carboxylase/smail subunit/time course/processing)

BERND REISS*, CATHERINE C. WASMANNt, JEFF SCHELL*, AND HANS J. BOHNERTt*Max-Planck-Institut fur Zuchtungsforschung, Egelspfad, D-5000 Koln 30 (Vogelsang), Federal Republic of Germany; and tDepartment of Biochemistry,University of Arizona, Tucson, AZ 85721

Contributed by Jeff Schell, September 26, 1988

ABSTRACT We studied transport and binding to intactchloroplasts of 10 mutants in three regions ofthe transit peptideofa precursor to the small subunit of ribulose 1,5-bisphosphatecarboxylase/oxygenase [3-phospho-D-glycerate carboxy-lyase(transphosphorylating), E.C. 4.1.1.39]. Transport was assayedin a reconstituted system using isolated pea chloroplasts andradioactively labeled precursor. Binding to the chloroplastenvelope was assayed in a similar manner using chloroplastspretreated with nigericin. Most mutants showed a dramaticallydecreased capacity of binding, although some of them trans-ported relatively well. The accumulation ofthe mutant proteinsinside the chloroplast as a function of time was examined.Although the authentic small subunit precursor was importedrapidly, uptake of most mutant precursors was considerablyslower and continued until the last time point examined. Interms of assigning functions to individual regions, we foundthat at least the middle region and parts of the amino and thecarboxyl termini of the transit peptide are more important forreceptor binding than for translocation. A two-step processingmechanism has been postulated for the maturation of the smallsubunit precursor. This model predicts the occurrence ofprocessing intermediates. When precursors carrying carboxyl-terminal deletions were presented to the chloroplast, no definedintermediates could be detected. Instead, a number of proteins,probably resulting from aberrant processing, accumulatedsimultaneously inside the chloroplasts.

The small subunit (SSU) of ribulose 1,5-bisphosphatecarboxylase/oxygenase [3-phospho-D-glycerate carboxy-lyase (transphosphorylating), E.C. 4.1.1.39] is one of theproteins that are synthesized in the cytoplasm and trans-ported posttranslationally into the chloroplast in an energy-dependent process (1, 2). All known proteins transported intothe chloroplast are synthesized as a precursor consisting ofthe mature protein and an amino-terminal extension, thetransit peptide (3-7). The transit peptide is proteolyticallyremoved upon transport and plays a major role not only inmediating import of the precursor into the correct organelle(8-12) but also in its subsequent localization within thechloroplast (13).The SSU precursor has been studied in great detail. From

results obtained with chemically modified SSU precursor orwith partially purified processing enzyme, a two-step matu-ration process has been postulated (14). Also, a SSU precur-sor of the green algae Chlamydomonas is imported by peachloroplasts and processed incompletely at a site apparentlyidentical with the postulated site of the first-stage processing(15). Many SSU genes from different organisms have beensequenced. By comparing the transit peptide sequences ofthese genes with those of other proteins transported into thechloroplast, three domains have been suggested as important

in chloroplast transport (2, 16). These domains are located atthe extreme amino and carboxyl termini and in the middleregion of the transit peptide. To test the significance of thesethree domains experimentally, we (11, 12, 17, 18) and others(19) introduced mutations into the transit peptide of SSUencoded by gene ss3.6 from pea, Pisum sativum (20). Weshowed that this transit peptide contains two different andseparated essential regions. Both the amino and the carboxyltermini were required for transport, whereas the middleregion, despite its evolutionary conservation, appeared toplay only a subordinate function.

Transport can be divided into three different theoreticalsteps, each involving the transit peptide. The precursor isdirected to the appropriate organelle and contacts a receptor(21, 22). Therefore, the first function must be a targeting andbinding function. Once the precursor has been bound, theprotein is translocated. Thus, the second function of thetransit peptide is aiding the precursor across the membrane.The third step in import ofthe SSU precursor is its processingto the mature form. The signal necessary for correct pro-cessing lies entirely within the carboxyl-terminal 13 aminoacids (18).

In this contribution we addressed the following questions.Can a specific function like binding, translocation, or pro-cessing be assigned to one of the regions defined earlier bydeletion-mutation analysis? Deletion mutations in the car-boxyl-terminal region of the transit peptide generate a seriesof incompletely processed polypeptides. Whether these ad-ditional polypeptides result from a multistep mechanism orrepresent the products of aberrant cleavage is unclear (17,18). To answer these questions, we analyzed the 10 mutantprecursors for their capacity to bind to the chloroplastsurface and for their transport competence as a function oftime in a reconstituted system.

MATERIALS AND METHODSPrecursor Synthesis. Synthetic mRNA was obtained from

described plasmids (17, 18) by transcription with SP6 poly-merase (23). Radioactively labeled precursor proteins wereobtained by translation of synthetic RNA in wheat germextracts in the presence of [35S]methionine (24).

Transport Experiments. Chloroplasts were isolated fromyoung pea plants Pisum sativum, variety Progress No. 9 andpurified on a Percoll gradient prepared in 1x grinding buffer(25) by high-speed centrifugation (26). The chloroplasts weresuspended in sorbitol/Hepes buffer (25) at a final concentra-tion of chlorophyll of 4 mg/ml; the chlorophyll concentrationwas determined according to Arnon (27).

Standart transport experiments were done as described(11, 25). A 300-,lI reaction contained 20 1.d of precursors inwheat germ extract and chloroplasts corresponding to 660 ,ug

Abbreviation: SSU, small subunit of ribulose 1,5-bisphosphate car-boxylase/oxygenase.

886

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 10

, 202

1

Proc. Natl. Acad. Sci. USA 86 (1989) 887

of chlorophyll per ml. The reaction mixture was incubated inlight at 20'C for 60 min. After incubation, the chloroplastswere collected by centrifugation and purified by repeatedwashing with sorbitol/Hepes buffer. The reaction mixturewas divided. One aliquot was protease-digested using ther-molysin as described (28) and then lysed; the other one waslysed directly.

Binding Assays. Binding assays were performed as de-scribed (28). Nigericin was added at a final concentration of400 nM to a standard transport reaction mixture from whichthe precursors were omitted. The mixture was preincubatedfor 10 min in the light, and then the precursors were added.After a 60-min incubation, samples were treated as describedfor the transport reactions with the exception that all solu-tions contained 400 nM nigericin.Time Kinetics. Individual transport reaction mixtures were

incubated in light for the time intervals given in Figs. 3 and4. The reactions were terminated by addition of ice-coldsorbitol/Hepes buffer containing 400 nM nigericin (28) andleft on ice. The chloroplasts were collected, purified, anddigested with thermolysin as described above.

Analysis of Proteins. Proteins were analyzed on 15%polyacrylamide gels containing NaDodSO4 (29). Radioac-tively labeled proteins were visualized by fluorography usingEN3HANCE (DuPont) following the manufacturer's instruc-tions. For quantification, autoradiograms of different expo-sure times were analyzed by microdensitometry. Transportand binding efficiencies, corrected for the different amount of[35S]methionine present in each individual precursor andprocessed protein, were calculated from that data.

APrecursors

1 2 3 4 5 6 7 8 9 10 11

pre-SSU > __ a D

0F w a

1B pssu

B T-

2 3 4 5 6iC6,"25 dli/5 dl/24 d6/25 d6/29

B T B _ T B T B T B _ Tf

T

Cf 8 9 1C0 11

d20I'35 d58 d42/5-8 d45/5f d48 5?

B T B T B T B T B T-f + _+ _ + _ +

RESULTSBinding Assays. The capacity of the various precursor pro-

teins to bind to the chloroplast surface was determined using thebinding assay developed by Cline et al. (28). Intact isolatedchloroplasts pretreated with the uncoupler nigericin lose thecapacity to import proteins. However, the precursors remainspecifically bound to the chloroplast surface. To determine boththe transport and the binding capacity of the same precursorpreparation, transport and binding were assayed in one and thesame experiment. To discriminate whether radioactively la-beled protein copurifying with the chloroplasts is located insidethe chloroplasts or attached to the organelle surface, proteaseprotection experiments were performed.The amino acid sequences of the authentic and of the 10

mutant precursors (17, 18) used in the experiments are shownin Fig. 1. Radioactively labeled precursor proteins wereobtained by SP6 polymerase transcription and translation inwheat germ extracts. Intact chloroplasts were isolated frompea, and the transport and binding experiments were per-formed as described. An aliquot ofeach reaction mixture wasprotease treated, and the precursor and chloroplast proteinswere analyzed on polyacrylamide gels containing NaDod-S04.The result of these experiments is shown in Fig. 2. The

Transit Peptide--------------------------------1 10 20 30 40

1 Pre-SSU MASMISSSAVTTVSRASRGQSAAVAPFGGLKSMTGFPVKKVNTDITS.

ISSSAVTTVSRASRGQSAAVA--MG___________-----------M

.I

.I

I.

I-----_------ |GI-------- G

a < pre-SSU

Ms< SSU7

FIG. 2. Effect of mutations on binding and transport. (A) Autora-diogram of the precursors. (B) Binding and transport experiments withauthentic SSU precursorand with amino-terminal mutants. (C) Bindingand transport of central and carboxyl-terminal mutants. The numbersof lanes in A correspond to panels in B and C. In B and C, B denotesthe binding assay, whereas T corresponds to the transport assay. -,Lanes of chloroplasts that are untreated; +, lanes of chloroplaststreated with thermolysin before lysis. The positions of the authenticSSU precursor and of the mature protein are indicated.

translation products obtained in the wheat germ systemcontained predominantly the desired precursor, althoughoccasionally additional proteins appeared. These extra pro-teins, which might result from abnormal initiation or termina-tion of transcription or translation, were neither bound to nortransported into chloroplasts to any significant extent andtherefore had no influence on the experiment. To quantitate theresults shown in Fig. 2, autoradiograms of these gels were

--------- Mature50 58

MQV

GIP QV-IQVMQVMQV

FIG. 1. Schematic representation ofthe transit peptide mutations. Theamino acid sequence of the transit pep-tide and the first three amino acids ofmature SSU are given in one lettercode. Deletions are represented bybars. The amino acid sequences arisingfrom insertions or from substitutionsare included. The numbering systemused for the mutants corresponds to theone used in Figs. 2 and 3.

c pre-SSU

< SSU

2 PSf6/253 PSdl1/54 PSdl/245 PSd6/256 PSd6/297 PSd26/358 PSd589 PSd42/5810 PSd45/5711 PSd48/57

Botany: Reiss et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 10

, 202

1

Proc. Natl. Acad. Sci. USA 86 (1989)

Table 1. Effect of mutations on binding and transport

Transport TransportMutant Binding - protease + protease

1 Pre-SSU 33 65 652 PSi6/25 36 77 773 PSd1/5 2 20 184 PSd1/24 1 0.5 0.55 PSd6/25 1 7 56 PSd6/29 1 5 37 PSd26/35 2 20 178 PSd58 31 69 409 PSd42/58 2 3 210 PSd45/57 1 7 611 PSd48/57 1 39 35

Binding and transport efficiencies are given in percentages ofprecursor offered to the chloroplasts. Values were calculated fromthe gels shown in Fig. 2.

analyzed by microdensitometry, and the transport and bindingefficiencies were calculated; these data are listed in Table 1.The import efficiency of the authentic SSU precursor

varied between 50% and 80% in different experiments. Theexact values for transport and binding efficiencies obtainedfor the mutants varied also between different experimentssimilarly. In a typical transport experiment, the authenticSSU precursor was imported to -65% by the chloroplasts.About 33% of the SSU precursor offered to the chloroplastsin the binding assay was recovered with the chloroplasts.Protease protection experiments revealed that this form isbound to the surface of the chloroplast because it wasdigested by externally added protease. The bound form of themutant precursors was also sensitive to protease in a similarfashion (data not shown).The different values obtained in the transport (65%) and in

the binding assays (33%) might appear puzzling; however,they simply reflect the experimental conditions chosen. In atransport experiment, several precursor molecules can passthe same receptor, whereas in the binding assay the precursoris bound to the receptor and therefore only as many precursormolecules as receptor molecules present in the assay can berecovered. Nigericin does not inhibit transport completely

(28). This is also evident from the slight amount of matureprotein visible in some lanes representing binding experi-ments (Fig. 2B pSSU and i6/25 lanes).Two of the 10 mutants examined, PSdi6/25 and PSd58, are

virtually identical to the authentic SSU precursor in theirbinding and transport capacities. The partial duplication ofthe transit peptide or the deletion of the first amino acid ofmature SSU, methionine, apparently does not interfere withany of these processes. All other mutants are impairedconsiderably in binding function. Mutants in which bothbinding and transport are severely affected and mutants thatbind weakly but transport relatively well can be distin-guished. Transport and binding are correlated, and both arerather inefficient with the amino-terminal mutants PSdl/24,PSd6/25, and PSd6/29 and with the two carboxyl-terminalmutants PSd42/58 and PSd45/57. The mutants that are boundpoorly but transported well represent relatively small dele-tions located either at the extreme amino or at the carboxylterminus or in the central region of the transit peptide.Removal of only five amino acids from the amino terminusreduces binding to almost the level of mutant PSd6/25.However, transport of this mutant is only moderately im-paired. The same applies to PSd26/35, although this deletionis nine amino acids in size. Particularly interesting is PSd48/57-this mutant is heavily affected in processing but not intransport. Nevertheless, this mutant has almost entirely lostits capacity to bind to the chloroplast surface.Time Kinetics. Intact pea chloroplasts were incubated with

radioactively labeled precursors in light for the time intervalsindicated in Fig. 3. The chloroplasts were recovered, washed,and protease-digested, and the chloroplast proteins wereanalyzed by polyacrylamide gel electrophoresis. Results areshown in Fig. 3; the data obtained by quantitation of thesegels are listed in Table 2.The authentic SSU precursor was almost completely taken

up by the chloroplasts after 5 min. In a shorter exposure thanthat shown, a slight increase between the 5- and 15-min timepoints was visible (data not shown). The imported andmatured SSU was fairly stable over the remaining timeperiod, because no appreciable decrease in signal strengthwas seen for up to 60 min. Two of the 10 mutants examined,

1pSSU

A 5 15 30 60

2 3i6/25 dl/5

5 15 30 60 5 15 30 60

4dl/24

5 15 30 60

5 6d6/25 d6/29

5 15 30 60 5 15 30 60

7d26/35

8d58

.r

..~~~~~~4w40.. _.., ;..i* A

9 10 11d42/58 d45/5 7 d48/5 7

B 5 15 30 60 5 15 30 60 5 15 30 60 5 15 30 60 5 15 30 60

<SSU

FIG. 3. Effect of mutations on the time course of import in the chloroplast. The autoradiogram shows the transport experiments performedwith the authentic SSU precursor and the amino-terminal (A) or central and carboxyl-terminal (B) mutants, respectively. Time intervals at whichthe samples were taken (5, 15, 30, and 60 min) are indicated.

888 Botany: Reiss et al.

040

Om 40 Oft

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 10

, 202

1

Proc. Natl. Acad. Sci. USA 86 (1989) 889

Table 2. Effect of mutations on the time course of import inthe chloroplast

Time, min

Mutant 5 15 30 60

1 Pre-SSU 54 62 60 542 PSi6/25 54 58 58 543 PSd1/5 2 6 10 134 PSd1/24 0 0 0 15 PSd6/25 0 0.5 1 26 PSd6/29 0 0.5 1 27 PSd26/35 2 5 8 118 PSd58 53 50 45 439 PSd42/58 0 2 2 410 PSd45/57 0 0.5 2.5 411 PSd48/57 11 26 38 30

Transport efficiencies are given in percentages of precursoroffered to the chloroplasts. Values were calculated from the gelsshown in Fig. 3.

PSi6/25 and PSd58, were indistinguishable from the authen-tic SSU precursor in their rate of transport into the chloro-plast. All other mutants showed various degrees of impair-ment.The larger amino-terminal (PSdl/24, PSd6/25, and PSd6/

29) and carboxyl-terminal (PSd42/58 and PSd45/57) deletionmutants were imported very slowly. After 5 min, virtually noimported protein could be detected. With PSdl/24, no im-ported SSU was visible before a 30-min incubation time. Withthe other four mutants, transport became detectable after 15min. Processed, mature SSU was already detectable at the5-min time point with mutants PSd1/5 (due to the low amount,not visible in Fig. 3) and PSd26/35, although the amounttransported during this time was considerably less than thatobtained with authentic SSU. The transported form of all theaforementioned mutants continued to accumulate constantlyinside the chloroplast up to 60 min.Mutant PSd48/57 was transported considerably faster than

all the other larger deletion mutants (Fig. 3B). An appreciableamount of product had already accumulated after 5 min, andtransport was almost completed at the 15-min time point. Thisindicates that the transport function in this transit peptide isnot severely affected by the mutation, although the mutationinterferes drastically with the processing function.Because the authentic SSU precursor and some mutants

are transported very rapidly, the time course experimentsdescribed above might not be suited to resolve subtle differ-ences in the initial rates of transport. Therefore, importexperiments using short time intervals were performed withthese mutants. The results of these experiments are shown inFig. 4.With the authentic SSU precursor, mature SSU was not

detected to a significant extent before the 2.5-min time point.At the 5-min time point, about twice the amount found after2.5 min was transported. Hence, these experimental condi-tions allow the determination of initial transport rates.PSi6/25 and PSd58 were imported as rapidly as the authenticSSU precursor. Therefore, these mutants are not evenmoderately impaired in their transport function. PSi6/25 mayeven be imported slightly faster than the authentic SSUprecursor. In this respect we note that the PSi6/25 transitpeptide leads to an enhanced import when connected to aforeign passenger protein (11). With mutant PSd48/57, trans-ported protein can be detected only after longer exposuretimes. As expected from the time course experiment shownin Fig. 3, transport of this mutant precursor was clearlyslower than that of the authentic SSU precursor. However,transport could already be detected after a 2.5-min incubationtime.

ApSSU i6/25 d58 d48/.57

0.5 1 2.5 5 0.5 1 2.5 5 0.5 1 2.5 5 0.5 1 2.5 5

<SSU

BpSSU i6/25 d58 d48/5 7

0.5 1 2.5 5 0.5 1 2.5 5 0.5 1 2.5 5 0.5 1 2.5 5_ __V_

<SSU.

FIG. 4. Effect of some mutations on the initial rate of import inthe chloroplast. Time intervals at which the samples were taken (0.5,1, 2.5, and 5 min) are indicated. (A) Exposure time, 1 day; (B)exposure time, 10 days.

The most prominent protein species detected after 0.5 or 1min of incubation migrated faster than mature SSU. Theseproteins were most likely derived from precursor that mayhave been bound to, or partially translocated into, thechloroplasts at the initial stages of transport and thereforecould not be completely digested in the protease treatmentstep. Similar protein species were also detected in the bindingexperiments after protease digestion.

All precursors carrying larger deletions at the carboxylterminus of the transit peptide were processed incorrectly.All products produced in addition to, or instead of, matureSSU appeared at all intervals during the time course exper-iments (see Discussion). With all mutants, no products otherthan the ones visible at 60 min are detected at any time.

DISCUSSIONWe have examined the binding and transport capacity of 10transit peptide mutants. Binding was assayed in intact iso-lated chloroplasts pretreated with the uncoupler nigericin. Ina parallel experiment, the transport capacity of the sameprecursor preparation was tested. In a typical experiment65% of the authentic SSU precursor was imported into thechloroplast, and -50% of the transported precursor wasbound to the chloroplast surface in a protease-sensitive formin the binding experiment. These values reflect the fact thatmore precursor molecules than receptor sites are presentunder the conditions used. To detect any processing inter-mediates and to determine the transport competence of theindividual precursors more precisely, time course experi-ments were done. The authentic SSU precursor nearlyreaches its maximum after =5 min. At the next time point, 15min, the reaction is complete, and the amount of mature SSUfound inside the chloroplast remains practically unchangeduntil the last examined point.According to their binding and transport characteristics,

the mutants can be grouped in three categories: (i) mutantsthat are unaffected, (ii) mutants that bind and transportinefficiently, and (iii) mutants that transport well but bindweakly.The mutants PSi6/25 and PSdS8 (category i) are indistin-

guishable from authentic SSU in their binding and transportcharacteristics. The PSi6/25 precursor contains a large du-plication of transit peptide sequences at the amino terminus

Botany: Reiss et aL

46lp40 db4,0% *&, , }:#:

a

4%

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 10

, 202

1

Proc. Natl. Acad. Sci. USA 86 (1989)

of the transit peptide. In the PSd58 precursor, the first aminoacid of mature SSU, methionine, is exchanged for thesequence Gly-Ile-Pro. Were a particular conformation of thetransit peptide necessary for function, both alterations wouldbe expected to interfere with transport or binding. The resultsobtained with these mutants, however, suggest that thetransit peptide tolerates fundamental structural changes.Therefore, it appears unlikely that the effects seen with thedeletion mutants are mainly due to a rather unspecificinterference with conformation. However, a clean distinctionbetween defined functional regions and effects of conforma-tion is impossible.The other mutants are all impaired in both transport and

binding, although to greatly different degrees. Mutations thatsimultaneously affect binding and transport (category ii) arelocated at the amino as well as at the carboxyl termini of thetransit peptide. The corresponding mutants are PSdl/24,PSd6/25, PSd6/29, PSd42/58, and PSd45/57, which are alltransported very inefficiently. Mutants PSd1/5 and PSd26/35are transported well but bind weakly. Mutant PSd48/57 istransported nearly as efficiently and only moderately slowerthan the authentic SSU precursor, but binding is virtuallyundetectable (category iii).We suggest that in the PSd48/57 transit peptide, the

mutation mainly affects the binding and processing functionswithout substantially interfering with the translocation func-tion. The time course of uptake seen with this mutant couldreflect a transport rate not dependent on a strong precursor-receptor interaction. All other mutants of category ii and iiiare severely affected in at least two functions-binding andtranslocation-with the exception of mutants PSd1/5 andPSd26/35, the translocation function ofwhich is less severelyaffected.How might precursors that bind poorly to chloroplasts be

nevertheless taken up rather efficiently? We offer two ex-planations. Possibly not only translocation (30) but alsobinding proceeds through several discrete steps, of whichonly the last one results in tight binding. If translocation canoccur without stabilization of binding and should nigericininhibit only this last step, we would see good translocationbut no binding. Alternatively, a strong precursor-receptorinteraction may be unnecessary in a reconstituted system.Under artificial conditions even very unstable interactionsbetween precursor and receptor may be sufficient for trans-location as soon as contact is made.

In conclusion, our results identify regions in the transitpeptide in which predominantly binding or receptor recog-nition is localized. These regions are the extreme ends andthe middle of the transit peptide. In addition, the carboxyl-terminal end of the transit peptide also contains a processingfunction. All other regions appear to operate both in bindingand translocation. No region exclusively involved in trans-location could be detected.The three carboxyl-terminal mutants PSd42/58, PSd45/57,

and PSd48/57 are affected in processing, as shown byabnormally cleaved products upon transport. However,whether these proteins result from a two-step processingmechanism (14) or are the products of aberrant processingremained to be answered. We addressed this question byfollowing the kinetics of import of these precursors. For atwo-step mechanism, the product of the first cleavage wouldbe predicted to appear first; unspecific degradation wouldoccur later. Therefore, a precursor-to-product relationshipshould be apparent in the time course experiments. Alterna-tively, in a one-step process involving aberrant cleavage, allthe imported proteins would be expected to appear simulta-neously inside the chloroplast. With all three carboxyl-terminal deletions, no precursor-to-product relationship was

detected in the time course-uptake experiments; instead, theproteins all appeared inside the chloroplast in a coordinatefashion. This then is the expected outcome if the SSUprecursor were processed in a single step and the additionalproducts resulted from aberrant cleavage. Therefore, wepropose that the alternative explanation, aberrant process-ing, has to be seriously considered.

We thank A. Blau and G. Simons for excellent technical assis-tance. C.C.W. and H.J.B. acknowledge support by the NationalScience Foundation (Grant PCM83-18166) and by Arizona Agricul-tural Experiment Station Grant ARZT-174452.

1. Ellis, R. J. (1981) Annu. Rev. Plant Physiol. 32, 111-137.2. Schmidt, G. W. & Mishkind, M. L. (1986) Annu. Rev. Bio-

chem. 55, 879-912.3. Chua, N.-H. & Schmidt, G. W. (1979)J. CellBiol. 81, 461-483.4. Schmith, S. M. & Ellis, R. J. (1979) Nature (London) 278, 662-

664.5. Grossman, A., Bartlett, S. & Chua, N.-H. (1980) Nature

(London) 285, 625-628.6. Smeekens, S., van Binsbergen, J. & Weisbeck, P. (1985)

Nucleic Acids Res. 13, 3179-3194.7. Della-Cioppa, G., Bauer, S. C., Klein, B. K., Shah, D. M.,

Fraley, R. T. & Kishore, G. M. (1986) Proc. NatI. Acad. Sci.USA 83, 6873-6877.

8. Schreier, P. H., Seftor, E. A., Schell, J. & Bohnert, H. J.(1985) EMBO J. 4, 25-32.

9. Van den Broeck, G., Timko, M. P., Kausch, A. P., Cashmore,A. R., Van Montagu, M. & Herrera-Estrella, L. (1985) Nature(London) 313, 358-363.

10. Lubben, T. H. & Keegstra, K. (1986) Proc. Natl. Acad. Sci.USA 83, 5502-5506.

11. Wasmann, C. C., Reiss, B., Bartlett, S. G. & Bohnert, H. J.(1986) Mol. Gen. Genet. 205, 446-453.

12. Kuntz, M., Simons, A., Schell, J. & Schreier, P. H. (1986) Mol.Gen. Genet. 205, 446-453.

13. Smeekens, S., Bauerle, C., Hageman, J., Keegstra, K. &Weisbeek, P. (1986) Cell 46, 365-375.

14. Robinson, C. & Ellis, R. J. (1984) Eur. J. Biochem. 142, 343-346.

15. Mishkind, M. L., Wessler, S. R. & Schmidt, G. W. (1985) J.Cell Biol. 100, 226-234.

16. Karlin-Neumann, G. & Tobin, E. M. (1986) EMBO J. 5, 9-13.17. Reiss, B., Wasmann, C. C. & Bohnert, H. J. (1987) Mol. Gen.

Genet. 209, 116-121.18. Wasmann, C. C., Reiss, B. & Bohnert, H. J. (1988) J. Biol.

Chem. 263, 617-619.19. Szabo, L. J. & Cashmore, A. R. (1987) in PlantDNA Infectious

Agents, eds. Hohn, T. H. & Schell, J. (Springer, Vienna), pp.321-329.

20. Cashmore, A. R. (1983) in Genetic Engineering of Plants: AnAgricultural Perspective, eds. Kosuge, T., Meredith, C. P. &Hollaender, A. (Plenum, New York), pp. 29-38.

21. Pain, D., Kanwar, Y. & Blobel, G. (1988) Nature (London) 331,232-237.

22. Cornwell, K. L. & Keegstra, K. (1987) Plant Physiol. 85, 780-785.

23. Krieg, P. A. & Melton, D. A. (1984) Nucleic Acids Res. 12,7057-7070.

24. Anderson, C. W., Straus, J. W. & Dudock, B. S. (1983) Meth-ods Enzymol. 101, 635-644.

25. Bartlett, S. G., grossman, A. R. & Chua, N.-H. (1982) inMethods i4 Chloroplast Molecular Biology, eds. Edelman, M.,Hallick, R. B. & Chua, N.-H. (Elsevier Biomedical, Amster-dam), pp. 1081-1102.

26. Cline, K., Andrews, J., Mersey, B., Newcomb, E. H. &Keegstra, K. (1981) Proc. NatI. Acad. Sci. USA 78, 3595-3599.

27. Amon, D. J. (1949) Plant Physiol. 24, 1-15.28. Cline, K., Werner-Washburne, M., Lubben, T. H. & Keegstra,

K. (1985) J. Biol. Chem. 260, 3691-3696.29. Laemmli, U. K. (1970) Nature (London) 227, 680-685.30. Singer, S. J., Maher, P. A. & Yaffe, M. P. (1987) Proc. Natl.

Acad. Sci. USA 84, 1015-1019.

890 Botany: Reiss et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 10

, 202

1