Title: Protein translocation on the Delta pH-dependent pathway of chloroplasts
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00100787/00001
 Material Information
Title: Protein translocation on the Delta pH-dependent pathway of chloroplasts
Physical Description: Book
Language: English
Creator: Fincher, Vivian
Publisher: University of Florida
Place of Publication: Gainesville Fla
Gainesville, Fla
Publication Date: 2001
Copyright Date: 2001
 Subjects
Subject: Plant Molecular and Cellular Biology thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Plant Molecular and Cellular Biology -- UF   ( lcsh )
Genre: government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )
 Notes
Summary: ABSTRACT: Plastids are evolutionarily derived from an endosymbiotic relationship with an ancient relative of cyanobacteria. The legacy of that endosymbiosis includes nuclear genes whose protein products must be imported into chloroplasts and routed to the thylakoid membranes or the lumenal space enclosed by those membranes. The Delta pH-dependent pathway is one of four chloroplast export pathways having prokaryotic ancestry. The Delta pH-dependent pathway is unique in its exclusive dependence on the trans-thylakoid pH gradient for energy, its use of transit peptides with a recognizable motif, and its ability to translocate folded proteins. I hypothesize that the Delta pH-dependent pathway has evolved from the most ancient means of protein translocation, spontaneous insertion. Early investigation of protein transport revealed initial interaction of amino acids with lipids and formation of a loop structure within the lipid bilayer. Subsequent research documented export initiation via loop formation in the Sec pathway of Escherichia coli and the endoplasmic reticulum. I have demonstrated that initiation of export of a fusion protein on the Delta pH-dependent pathway also occurs via loop formation. Evolutionary advances in control of protein translocation have resulted in machinery that can monitor substrates for appropriate characteristics such as targeting sequences, folding, and cofactor insertion. When a substrate does not meet criteria for continued translocation, transport-arrest results.
Summary: ABSTRACT (cont.): I have captured a membrane-spanning translocation intermediate resulting from transport-arrest on the Delta pH-dependent pathway. I characterized the interaction by mutation of the substrate-protein and by chemical treatments of the recovered membranes. Manipulation of the intermediate may allow investigation of the transport process and of translocation machinery. The known components of the Delta pH-dependent pathway machinery comprise three membrane-associated proteins. I have demonstrated incorporation of in vitro-translated components into complexes formed by endogenous components visualized by blue native polyacrylamide gel electrophoresis and fluorography. Incorporation of in vitro-translated and radiolabeled components allows tracking of substrate and component interactions. I have documented that the change of a single amino acid in one component can abolish the incorporation of that component into endogenous complexes. The methodology developed in this project will be used in subsequent research on component and substrate interactions.
Summary: KEYWORDS: chloroplast, protein transport, Tat, thylakoid, translocation
Thesis: Thesis (Ph. D.)--University of Florida, 2001.
Bibliography: Includes bibliographical references (p. 92-97).
System Details: System requirements: World Wide Web browser and PDF reader.
System Details: Mode of access: World Wide Web.
Statement of Responsibility: by Vivian Fincher.
General Note: Title from first page of PDF file.
General Note: Document formatted into pages; contains ix, 98 p.; also contains graphics.
General Note: Vita.
 Record Information
Bibliographic ID: UF00100787
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 50743562
alephbibnum - 002729345
notis - ANK7109

Downloads

This item has the following downloads:

vivian ( PDF )


Full Text











PROTEIN TRANSLOCATION ON THE DELTA pH-DEPENDENT PATHWAY OF
CHLOROPLASTS

















By

VIVIAN FINCHER


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2001




























Copyright 2001

by

Vivian Fincher




















This dissertation is dedicated to those few men and women who strive to exercise
genuine integrity in the pursuit of science and to those less perfect humans who must
admire and endure them. While genuine integrity is the underpinning of all science, the
burden it imposes on the scientists is a heavy one. By its very definition, integrity makes
no exceptions, is indifferent in the face of desire and unrelenting in the wake of fatigue.















ACKNOWLEDGMENTS


The author wishes to thank her advisor, Dr. Kenneth C. Cline, for his exceptional

accessibility and guidance throughout her research project. Dr. Cline has provided a

consistent model of integrity and a display of work ethic that have been invaluable. She

also expresses gratitude to her other committee members, Dr. Robert J. Ferl, Dr. Alice C.

Harmon, Dr. Paul C. Sehnke and Dr. John P. Aris, for their patience and scientific

support. An additional acknowledgement goes to Dr. Ralph Henry who spent many

hours instructing her in laboratory techniques. This work would not have been possible

without the technical advice and assistance of Mike McCaffery and Dr. Hiroki Mori.
















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ................................................................................................. iv

A B S T R A C T ........................................................................................................ v iii

CHAPTERS

1 LITER A TU R E R EV IEW .................................................... ............................... .. 1

Modem Chloroplasts Are Descended from Photosynthetic Prokaryotes. .................... 1
The Bacterial Tat System Is Evolutionarily Related to the Chloroplast
ApH -dependent Pathw ay ................................................................................. .... 4
The Chloroplast ApH-dependent Pathway Was Explored First Biochemically, Then
Genetically. ............................................. 10
Sum m ary and Perspective ............................................................ .............. 17

2 EVIDENCE FOR A LOOP MECHANISM OF PROTEIN TRANSPORT BY THE
THYLAKOID ApH-DEPENDENT PATHWAY ..................................................19

In tro d u ctio n .................................................................................... 19
M materials and M ethods................... ............................................. .................... ...... 21
M materials .............................. ....... ............... .......... 21
Construction of Chim eric Precursors........................ ........................ ............. 21
Preparation of Radiolabeled Precursors by In Vitro Translation............................. 22
Preparation of Chloroplasts, Thylakoids, and Lysate............................................. 23
Chloroplast Import and Thylakoid Protein Transport Assays .................................. 23
Im m unoprecipitation ........................................................ ................ .. ....... 24
A analysis of Sam ples ........................................................ ... . .. ...... ..... 24
R esults.......................... ....... ... ....... ... ............................... . 25
A Chimeric Precursor Protein for Examination of Transport Topology Was
C constructed ........................... ..... ........ ................................... 25
m23p Is Localized to the Cis Side of the Membrane, whereas mOE17 Is in the
L u m en ................ .................... .......... ...... ........... ................ ............... 2 8
Production of m23p and mOE17 from m23pl7 Results from ApH-dependent
P athw ay T transport ........................................... ................... ..... 29
D iscu ssio n ...................................... ..................................... ............... 3 1









3 THE CHARACTERIZATION OF A TRANSLOCATION INTERMEDIATE ON
THE ApH-DEPENDENT PATHW AY ..................................................................... 35

In tro d u ctio n .................................................................................... 3 5
Materials and Methods....................... ......... ............... .............. 36
Preparation of Chloroplasts, Thylakoids, and Lysate............................................. 36
Construction of Chim eric Precursors.................................................................... 37
Preparation of Radiolabeled Precursors................................................................. 41
Chloroplast Import and Thylakoid Protein Transport Assays ................................. 42
Im m unoprecipitation...................................................... .......... .... .. ......... 42
R esults.......................... ..... . ........... .. .... .... ......... ..... ............... . 43
Fusion Proteins Consisting of pOE17 and Protein A Were Constructed Having
V ary in g L in k ers ........................... ...... ............ ......... ........................ ........ 4 3
Fractionation of Chloroplasts Following Import Demonstrated a Thylakoid
Localized M ature Form of 17-protA ............................................................. 45
Processing of p 17-protA to the Mature Size Occurs Only under Transport
Perm issive Conditions ................................................. ................ ................ 47
Following Processing by the Lumenal Protease, the Mature Substrate Spans the
Thylakoid Membrane with Its Amino Terminus on the Trans Side of the
Membrane and Its Carboxyl Terminus on the Cis Side of the Membrane.......... 48
The Membrane-spanning Intermediate Is Not Integrated into the Lipid Bilayer. .... 53
The Membrane-spanning Intermediate is Arrested Due to the Protein A Moiety.... 56
The Protein A Moiety Alone Is Insufficient to Arrest Transport............................. 59
D isc u ssio n ..................................................................................... 6 0

4 MEMBRANE INTEGRATION OF IN VITRO-TRANSLATED
ApH-DEPENDENT PATHWAY COMPONENTS ...............................................63

In tro d u ctio n ......................................................... ............... 6 3
M materials and M ethods................... ............................................. .................... ...... 65
Preparation of Precursor Proteins .................................. ............................. ........ 65
Preparation of Radiolabeled Precursors........................... ......... ............. 67
Preparation of Chloroplasts, Thylakoids, and Lysate............................................. 67
Chloroplast Import and Thylakoid Protein Integration Assays ............................. 68
Q uantitative Im m unoblots ................................................... ................... ...... 68
Blue Native Polyacrylamide Gel Electrophoresis.................................. 69
R esults............................ ....................... .. ....... .................... . 69
In Vitro-translated ApH-dependent Pathway Components Associated with
Thylakoid M em branes ..................................... .... .. .... .................. 69
Association of In Vitro-translated Components with Endogenous Complexes........ 75
Integration of In Vitro-translated Components Influenced Efficiency of
Subsequent ApH-dependent Pathway Transport. .......................................... 77
Investigation of the Interaction of In Vitro-translated Components with Maize
Membranes. ........................................... 80
D iscu ssio n ...................... .. .. ......... .. .. ................................................ 8 5









5 SUMMARY AND CONCLUSIONS ........................................ ........................ 88

L IST O F R E FE R E N C E S .................................................. ........................ ....................92

B IO G R A PH IC A L SK E T C H ............................... ................. ........................................98















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

PROTEIN TRANSLOCATION ON THE DELTA pH-DEPENDENT PATHWAY OF
CHLOROPLASTS

By

Vivian Fincher

May 2001

Chairman: Kenneth C. Cline
Major Program: Plant Molecular and Cellular Biology

Plastids are evolutionarily derived from an endosymbiotic relationship with an

ancient relative of cyanobacteria. The legacy of that endosymbiosis includes nuclear

genes whose protein products must be imported into chloroplasts and routed to the

thylakoid membranes or the lumenal space enclosed by those membranes. The Delta pH-

dependent pathway is one of four chloroplast export pathways having prokaryotic

ancestry. The Delta pH-dependent pathway is unique in its exclusive dependence on the

trans-thylakoid pH gradient for energy, its use of transit peptides with a recognizable

motif, and its ability to translocate folded proteins.

I hypothesize that the Delta pH-dependent pathway has evolved from the most

ancient means of protein translocation, spontaneous insertion. Early investigation of

protein transport revealed initial interaction of amino acids with lipids and formation of a

loop structure within the lipid bilayer. Subsequent research documented export initiation

via loop formation in the Sec pathway of Escherichia coli and the endoplasmic reticulum.









I have demonstrated that initiation of export of a fusion protein on the Delta pH-

dependent pathway also occurs via loop formation.

Evolutionary advances in control of protein translocation have resulted in

machinery that can monitor substrates for appropriate characteristics such as targeting

sequences, folding, and cofactor insertion. When a substrate does not meet criteria for

continued translocation, transport-arrest results. I have captured a membrane-spanning

translocation intermediate resulting from transport-arrest on the Delta pH-dependent

pathway. I characterized the interaction by mutation of the substrate-protein and by

chemical treatments of the recovered membranes. Manipulation of the intermediate may

allow investigation of the transport process and of translocation machinery.

The known components of the Delta pH-dependent pathway machinery comprise

three membrane-associated proteins. I have demonstrated incorporation of in vitro-

translated components into complexes formed by endogenous components visualized by

blue native polyacrylamide gel electrophoresis and fluorography. Incorporation of in

vitro-translated and radiolabeled components allows tracking of substrate and component

interactions. I have documented that the change of a single amino acid in one component

can abolish the incorporation of that component into endogenous complexes. The

methodology developed in this project will be used in subsequent research on component

and substrate interactions.















CHAPTER 1
LITERATURE REVIEW


Modern Chloroplasts Are Descended from Photosynthetic Prokaryotes.

While the precise nature of the endosymbiotic events leading to the formulation of

modern eukaryotic cells is a subject of intense debate, the derivation of modern

chloroplasts from an ancient relative of cyanobacteria is well accepted. The prokaryotic

ancestry of chloroplasts is evidenced genetically, topologically and biochemically.

Chloroplasts and, more generally speaking, all plastid types found in various

developmental and functional forms of plant cells have retained portions of the bacterial

chromosome. Genes encoded in the plastid genome are transcribed and translated within

the plastid using machinery that is formulated from a mix of nuclear and plastid-encoded

proteins and RNA. However, the majority of formerly bacterial genes has been

transferred to the plant nucleus. The integration of plastid-gene and nuclear gene

products into a multi-compartmented functional organelle is dependent on the plastid

envelope import machinery and the thylakoid membrane export machinery.

The topological relationships of the compartments of the chloroplast reflect its

prokaryotic heritage. The organelle is delimitated by a double membrane comprising the

outer envelope, inter-envelope space, and inner envelope. Within the organelle, the fluid

compartment called the stroma is analogous to the bacterial cytoplasm. The chloroplast

has an extensive intra-organellar membrane system termed the thylakoid membranes.

Within the thylakoid membranes and the lumen they enclose resides the chloroplast









photosynthetic machinery. The thylakoids are derived from the bacterial inner membrane

and thus their enclosed space, the thylakoid lumen, is topologically equivalent to the

bacterial periplasmic space. Therefore, export processes from the stroma to the thylakoid

membrane or lumen accomplish translocation of the proteinaceous components of the

photosynthetic machinery (Schatz and Dobberstein, 1996).

Targeting of cytoplasmically translated proteins to the chloroplast stroma or the

thylakoid compartments is controlled by cleavable N-terminal transit peptides (Keegstra

and Cline, 1999). For proteins destined to the stroma and for many thylakoid membrane

proteins, the transit peptides consist of a single domain, which is both necessary and

sufficient for transport of the precursors across the chloroplast envelope. Integral

membrane proteins contain additional targeting information, typically coded in their

membrane-spanning domains. Those transit peptides of precursors destined for the

thylakoid lumen, as well as some membrane-targeted proteins, have a lumenal targeting

domain following the cleavage site of the stromal-targeting domain. The sequence and

structure of the lumenal targeting domain determines the export pathway utilized by the

protein. Following translocation of the stromal intermediate, the lumenal targeting

domain is cleaved by the lumenal protease resulting in the mature sized protein.

Investigation of energetic and stromal component requirements for protein export has led

to the defining of four translocation pathways. One group of integral thylakoid

membrane proteins may be inserted spontaneously without the aid of proteinaceous

components. Three other groups are dependent on complex membranal, and in some

cases stromal, machinery.









A subset of thylakoid lumenal proteins is translocated by the chloroplast Sec

(cpSec) pathway (Keegstra and Cline, 1999). The lumenal targeting domains of Sec

pathway precursors are characterized by a tripartite signal peptide structurally similar to

prokaryotic Sec system signal peptides. The two components of the pathway that have

been identified genetically, cpSecY, and cpSecA, are homologous to bacterial Sec system

proteins. Purified cpSecA has been shown to promote ATP dependent and azide

sensitive translocation of the 33-kDa subunit of photosystem II oxygen evolving complex

(OE33) and plastocyanin (PC). cpSec pathway directed proteins interact with cpSecA on

the thylakoid membrane forming a stable complex in the absence of ATP. Following

addition of ATP, a portion of the complex bound precursor can be exported to the lumen.

The degree to which export is stimulated by the trans-thylakoid pH gradient (ApH) is

substrate dependent. Translocation can be blocked by antibody bound to cpSecY (Mori

et al., 1999) and can be competitively inhibited only by other cpSec pathway precursors.

Export of a family of nuclear encoded integral thylakoid proteins, the light

harvesting chlorophyll-binding proteins (LHCP), is dependent on the chloroplast signal

recognition particle-like pathway (Keegstra and Cline, 1999). Following cleavage of its

stromal targeting domain, LHCP interacts with the chloroplast homologue of the

mammalian SRP54 subunit (also homologous to Escherichia coli Ffh) and a novel 43-kD

protein, cpSRP43, to form a 120-kD soluble transit complex. Additionally, GTP and the

stromal protein cpFtsY, chloroplast homologue of mammalian SRP receptor c-subunit

(and E. coli FtsY), have been demonstrated to be essential for LHCP integration (Kogata

et al., 1999). The trans-thylakoid pH gradient is stimulatory, but not essential for LHCP

integration. Integration of LHCP into the thylakoid membrane is dependent on Oxalp,









the chloroplast homologue of mitochondrial Oxalp (and E. coli YidC), and independent

of cpSecY (Moore et al., 2000). Integration can be competitively inhibited by LHCP, but

not by cpSec pathway or ApH-dependent pathway directed precursors.

Export of proteins to the thylakoid lumen via the ApH-dependent pathway is

energetically unique in its sole dependence on the trans-thylakoid pH gradient (Keegstra

and Cline, 1999). Proteins targeted to this pathway are synthesized with a signal peptide

structurally similar to bacterial Sec transit sequences, but characterized by a twin arginine

motif immediately preceding their hydrophobic domain. Membrane localized

components of the ApH-dependent pathway were first characterized genetically in plants

and later in prokaryotes where the system is designated the Tat (for twin-arginine

translocation) pathway. As the focus of this dissertation, the ApH-dependent pathway

will be discussed in considerable detail.


The Bacterial Tat System Is Evolutionarily Related to the Chloroplast ApH-
dependent Pathway.

The primary export system in prokaryotes is the general secretary (Sec) pathway.

Translocation via the Sec pathway is mediated by the peripheral ATPase SecA and

membrane components including the SecYEG complex (Manting and Driessen, 2000).

Among the auxiliary proteins of the Sec system are SecD and SecF, which are essential to

maintenance of the proton motive force (Arkowitz and Wickner, 1994) and regulate SecA

activity (Economou et al., 1995). It has been observed that fully sequenced archaeal

genomes encode either homologues of SecD and SecF or homologues to Tat pathway

proteins TatC and Hcfl06, but not both (Eichler, 2000). Since homologues for SecA









were not found among archaeal primary sequences, the role of SecD and SecF in archaea

is unknown.

Transport by the bacterial Sec pathway is initiated by insertion of the preprotein

into the membrane as a loop (Kuhn et al., 1994) and then progresses by linear

translocation of the extended protein chain through the membrane from its amino end to

carboxyl end. Because the bacterial Sec machinery requires its substrate proteins to

adopt an extended conformation during translocation (Pugsley, 1993), it is unsuitable for

export of proteins or protein complexes that must be assembled and tightly folded in the

cytoplasm. Prominent among proteins acquiring tightly folded conformations while still

cytoplasmically located are the metallo-enzymes and nucleotide cofactor containing

enzymes required for anaerobic respiration (Berks et al., 2000a). A survey of the E. coli

genome revealed twenty-three open reading frames encoding proteins having signal

peptides containing a Tat consensus sequence. Of those twenty-three precursors, sixteen

bind or are predicted to bind redox factors (Stanley et al., 2001). Assembly of many

apoprotein-cofactor complexes is a cytoplasmic process involving dedicated assembly

factors and proteases. Indirect evidence from numerous in vivo experiments supports a

model in which fully folded complexes or oligomers are translocated via the Tat

apparatus (Berks et al., 2000b, for review). When green fluorescent protein (GFP) was

fused behind the premaltose-binding protein and expressed in E. coli, it was exported via

the Sec pathway (Feilmeier et al., 2000). The chimeric construct was fully fluorescent in

the cytoplasm, but not in the periplasm. The authors concluded that GFP cannot fold

correctly in the periplasm. A similar construct having a Tat system compatible signal









peptide was fluorescent in both compartments, leading to the conclusion that it could be

exported in its folded state (Thomas et al., 2001).

Although Tat signal sequences share overall structure with Sec signal sequences,

some differences are important. Both sets of peptides can be described as tripartite,

having a basic N-region at the amino terminus followed by a hydrophobic H-region and

ending in a carboxyl C-region containing the signal peptide cleavage site. The N-region

of Tat precursors is on average 14 amino acids longer than the corresponding domain in

Sec signal peptides with most of that difference being accounted for by the characteristic

twin arginine consensus region (Crist6bal et al., 1999). Berks (1996) defined the

bacterial twin arginine consensus as (S/T)-R-R-X-F-L-K immediately preceding the H-

region. While the two consecutive arginines are invariant, the other amino acids of the

consensus may differ in about half of Tat signal peptides. Multiple experimental studies

have confirmed that the of the twin arginine motif is essential for export (Berks et al.,

2000b, for review). Positive charges are more common in the C-region of Tat signal

peptides than in the corresponding domain of Sec precursors (Crist6bal et al., 1999).

Comparing a series of chimeric constructs between variations on the signal

peptide of trimethylamine N-oxide reductase (TorA) and the periplasmic P2 domain of E.

coli protein leader peptidase (Lep) led Crist6bal et al. (1999) to conclude that the overall

lower hydrophobicity of the H-region of Tat peptides is important both in Sec avoidance

and in Tat transport. Their conclusion was based on data that demonstrated exclusive

transport on the Sec pathway when the TorA signal peptide was mutated to contain a

more hydrophobic H-region still flanked by its native N- and C- regions. Transport in tat

deletion mutants was unaffected, while depletion of SecE completely eliminated









translocation. Izard and Kendall (1994) suggested that the apparent decrease in

sensitivity to sodium azide inhibition of transport in Sec precursors with artificially

increased hydrophobic domains might reflect more efficient utilization of SecA. If the

increase in hydrophobicity of the mutated TorA signal peptide is increasing affinity for

SecA, sequestering of the precursor may be responsible for the apparent inability of the

Tat system to translocate the protein in the SecE depleted cells. Similar constructs

having Sec type hydrophobic domains but carrying twin arginine consensus sequence are

capable of utilizing either the cpSec or ApH-dependent pathway in chloroplasts (Henry et

al., 1997). The increased hydrophobicity of the modified TorA H-domain may be

necessary to allow Sec dependent transport, but may not inherently prohibit Tat pathway

transport. A similar experiment in SecA depleted cells might determine the answer.

The interaction of Tat machinery and E. coli membranes has received recent

attention. Mikhaleva et al. (1999) examined the role of phospholipids in the translocation

of TorA. They concluded that the Tat pathway is more highly dependent on non-bilayer

preferring phospholipids, such as phosphatidylethanolamine, than is the Sec pathway.

They suggested, "the slow process of translocating folded protein through the Tat

pathway may increase the opportunity for a direct, active interaction between the

phospholipids and the passenger proteins" (Mikhaleva et al., 1999 p. 335). Stanley et al.

(2001) have observed pleiotropic defects in the cell envelope of several E. coli mutant

strains blocked in Tat translocation. The bacteria appear to be defective in cell

separation, forming chains of up to 10 cells. The cells, which are resistant to infection by

P1 phage, are supersensitive to killing by hydrophobic drugs and to lysis by lysozyme in









the absence of EDTA. The phenotype was construed to be due to a defect in the

biosynthesis of the outer membrane.

The genetic definition of the Tat pathway began with the isolation of a maize

mutant termed hcfl06. Chloroplasts from hcfl06 plants are selectively deficient in the

export of ApH-dependent pathway precursors OE17 and OE23 (the 17-kDa and the 23-

kDa subunits of the photosystem II oxygen-evolving complex, respectively), but are

normal in the targeting of cpSec dependent precursors OE33 and PC (Voelker and

Barkan, 1995). Cloning of the Hcfl06 gene and comparison of its sequence to available

bacterial genomes led Settles et al. (1997) to conclude that the thylakoid ApH-dependent

pathway evolved from a bacterial redox protein secretary system. In E. coli the primary

operon controlling that system had been designated yigTUW.

Formerly designated yigTUW, the Tat operon located at 86 minutes on the E. coli

chromosome was analyzed, resequenced, and redefined to contain four genes TatA, TatB,

TatC, and Tat D (Weiner et al., 1998; Sargent et al., 1998). An independent

transcriptional unit, TatE is located at 14 minutes (Sargent et al., 1998). The TatA, TatB

and TatE genes code for homologues of the maize Hcfl06 protein. Each protein

comprises an N-terminal transmembrane a-helix followed by an amphipathic a-helix

and, in the case of TatB, an extended C-terminal region. Analysis of sequence data from

all complete genomes of bacteria has shown that homologues of TatA/B/E and of TatC

are present in all organisms that encode proteins with twin-arginine transit peptides

(Berks et al., 2000b). Additionally, TatC homologues are found in the mitochondrial

genomes of Arabidopsis thaliana, Marchantia polymorpha, and Reclinomonas americana

(Bogsch et al., 1998). TatC is predicted to span the membrane six times with amino and









carboxyl termini in the cytoplasm (Sargent et al., 1998). TatD related sequences are

found in all complete genomes except that ofArchaeoglobusfulgidus (Wu et al., 2000);

however, it is not conserved in linkage to other Tat genes (Settles and Martienssen,

1998).

Manipulation of Tat genes in E. coli has allowed the roles of the gene products to

be assessed separately and together. It was found that mutation (Weiner et al., 1998) or

deletion (Sargent et al., 1999) of the TatB gene eliminated transport of Tat pathway

dependent precursors. In a pulse-chase experiment, TatC protein was shown to be

unstable in the absence of TatB, implying an interaction of the gene products (Sargent et

al., 1999). Western blots of a TatA deletion strain revealed only very low levels of TatB,

suggesting that TatB is unstable in the absence of TatA (Bolhuis et al., 2000). While

disruption of TatA or TatE individually impaired but did not eliminate Tat dependent

protein export, deletion of both the TatA and TatE genes completely blocked

translocation (Sargent et al., 1998). Co-immunoprecipitation experiments in E. coli

demonstrated that about 5% of TatA is associated with TatB (Bolhuis et al., 2000). As

part of the same study, gel filtration results demonstrated a 600- kDa complex containing

both TatA and TatB. Disruption of the TatC gene also blocked Tat pathway dependent

export (Bogsch et al., 1998). In E. coli the cytoplasmic protein, TatD, is not required for

translocation of Tat system dependent proteins (Wexler et al., 2000).

Purification of the TatA/TatB/TatC complex revealed a 1:1 stoichiometry of TatB

to TatC with TatA present in substoichiometric amounts (Bolhuis et al., 2001). The

authors concluded that TatA is more loosely associated to the complex than TatB/TatC.









Co-immununoprecipitation results suggested that TatC is required for the interaction of

TatA and TatB.

The independence of the Tat and Sec systems in E. coli has been demonstrated in

several experiments. Neither disruption of TatB (Weiner et al., 1998), deletion of both

TatA and TatE (Sargent et al., 1998) nor deletion of TatC (Bogsch et al., 1998) had

discernable effects on export via the Sec pathway. Conversely, Crist6bal et al. (1999)

found that SecE depletion had no effect on the translocation kinetics of the Tat pathway

directed TorA/P2 fusion protein.


The Chloroplast ApH-dependent Pathway Was Explored First Biochemically, Then
Genetically.

Definition of the ApH-dependent pathway in chloroplasts was achieved first

through biochemical experiments. Thylakoids that have been recovered from lysed

chloroplasts and subsequently washed and resuspended in buffer are capable of

translocating ApH-dependent precursors when provided with actinic light at 250 C.

Rigorous experimentation has led to the consensus that the pathway requires no

nucleotide triphosphates (NTP's) and no soluble stromal components, in contrast to both

the cpSec and cpSRP pathways (Mould and Robinson, 1991; Cline et al., 1992; Cline et

al., 1993; Hulford et al., 1994). Translocation is energically dependent on the trans-

thylakoid pH gradient. Where tested, export of any ApH-dependent pathway precursor

can be competitively inhibited by any other ApH-dependent pathway precursor, but not

by cpSec or cpSRP pathway precursors (Cline et al., 1993).

The structure of the signal peptides of ApH-dependent precursors is similar to the

structure of bacterial Tat precursors, cpSec precursors, and bacterial Sec precursors. All









share the tripartite structure described above. Invariably, ApH-dependent precursors bear

two arginine residues preceding the H-domain. Domain swapping experiments among

precursors of the cpSec and ApH-dependent pathways have demonstrated that the two

arginines are required for ApH-dependent targeting of lumenal proteins (Chaddock et al.,

1995; Henry et al., 1997) and that the degree of influence of the H-domain and C-domain

amino acids varies among precursors (Bogsch et al., 1997; Henry et al., 1997). Signal

peptides that direct a protein to the cpSec pathway are unable to promote export of

proteins naturally transported by the ApH-dependent pathway (Clausmeyer et al., 1993;

Henry et al., 1997). However, signal peptides directing export by the ApH-dependent

pathway are able to promote transport of plastocyanin (PC), which is naturally exported

by the cpSec pathway (Robinson et al., 1994; Henry et al., 1994; Henry et al., 1997). A

signal peptide comprising the N-domain of OE23, the H-domain of PC, and the C-

domain of PC is able to direct PC and OE33 to both the cpSec and the ApH-dependent

pathways (termed dual-targeting); the same signal peptide directs OE23 exclusively on

the ApH-dependent pathway (Henry et al., 1997). Henry et al. (1997) hypothesized that

the ApH-dependent pathway has evolved to export precursors that the cpSec pathway is

unable to translocate.

The identification of the hcfl06 mutant in maize and the demonstration that it is

deficient specifically in ApH-dependent pathway precursor export (Voelker and Barkan,

1995) lay the foundation for system component identification and study. Subsequently,

the Hcfl06 gene was cloned and homologous sequences were found in prokaryotic

genomes (Settles et al., 1997). Further analysis in maize revealed Tha4 and Tha9 as

orthologues of E. coli TatA. Hcfl06 and a gene bearing >90% identity to it were









described as orthologues to E. coli TatB (Walker et al., 1999). Additional evidence of the

common evolutionary origin of the bacterial Tat pathway and the chloroplast ApH-

dependent pathway came from experiments exploring the compatibility of bacterial signal

peptides with thylakoid export machinery. Mori and Cline (1998) fused the signal

peptide from E. coli hydrogenase 1 small subunit to the mature domain of plastocyanin

and demonstrated that isolated thylakoids transport it by the ApH-dependent pathway.

Glucose-fructose oxidoreductase from Zymomonas mobilis is similarly exported by

isolated thylakoids (Halbig et al., 1999).

The biochemical independence of the cpSec and ApH-dependent pathways has

been demonstrated by multiple methods from several research groups. Early evidence of

separate translocation systems came from competition studies as mentioned previously.

As a bridge between the genetic system of maize and the biochemical system of pea,

Tha4 (Mori et al., 1999), Hcfl06 and cpTatC (Mori and Cline, unpublished) were cloned

from Pisum sativum. Following the cloning of pathway components, antibodies to those

components were tested for in vitro inhibition of export. When isolated thylakoid

membranes are pre-treated with antibodies to Tha4, Hcfl06, or cpTatC, only

translocation of ApH-dependent pathway substrates is inhibited (Mori et al., 1999;

Summer et al., 2000). Pretreatment of membranes with antibody to cpSecY inhibits only

cpSec dependent precursors (Mori et al., 1999).

Independent evidence that cpSec and ApH-dependent pathway precursors use

different translocation systems came from Asai et al. (1999). A fusion protein

comprising the stromal intermediate of OE23 (iOE23) and E. coli biotin carboxyl carrier

protein (i23K-BCCP) was expressed in E. coli. Purified biotinylated i23K-BCCP is









translocated to the lumen of isolated thylakoids where it is processed to mature size

(m23K-BCCP) and protected from externally added thermolysin. Transport is dependent

on the ApH and is azide insensitive as expected for a ApH-dependent pathway precursor.

When i23K-BCCP is completed with avidin prior to incubation with isolated thylakoids,

the mature sized product is formed indicating processing by the lumenal protease;

however, the m23K-BCCP remains sensitive to thermolysin. These results imply that

export has been arrested with the amino end of the substrate protein in the lumen while

most of the protein remains on the cis side of the membrane. When large amounts of

avidin completed i23K-BCCP are incubated with thylakoids, the transport of ApH-

dependent pathway stromal intermediate iOE23 is blocked. The membranes are still

competent to transport cpSec directed stromal intermediate iOE33. The authors

concluded, "this means that the translocation pore for the ApH-dependent pathway is not

shared by substrates for the Sec-dependent pathway" (Asai et al., 1999, p. 20077).

Noteworthy also is the implication that biotinylated iOE23-BCCP bound with avidin is an

unacceptable substrate for complete translocation due either to size or some other feature

of the complex. When a similar construct, pOE17(C)-BioHis, used by Musser and Theg

(2000a) is biotinylated and completed with avidin, it binds to the membrane and

competitively inhibits translocation of unbiotinylated precursor. However, the avidin

completed protein is not processed to mature size. The reason for the difference in

behavior of precursors in the two studies remains unexplained.

The most unique feature of the ApH-dependent system is its ability to translocate

folded proteins. The first intimation of folded protein transport by the pathway came

from a 1995 study in which Creighton et al. demonstrated that iOE23 contains a protease









resistant, and by implication tightly folded, 22-kDa core. Because data from other

translocation systems indicate that proteins are typically transported in an unfolded state,

the authors concluded that OE23 was probably unfolded during export. Subsequently,

Clark and Theg (1997) demonstrated that the tightly folded 6.5-kDa bovine pancreatic

trypsin inhibitor (BPTI) linked behind the ApH-dependent pathway precursor to OE17

(pOE17) could be fully translocated to the thylakoid lumen. Transport proceeds even

when the BPTI moiety, estimated to be 2.3 nm in diameter, is internally crosslinked and

therefore incapable of unfolding during translocation. The export of proteins is achieved

without compromising the ability of the thylakoids to maintain an ion gradient (Teter and

Theg, 1998). Evidence of the system's ability to export a somewhat larger folded domain

came from a study in which dihydrofolate reductase (DHFR) was linked to pOE23 and

shown to enter the lumen even when bound to methotrexate (Hynds et al., 1998). In that

series of experiments, it was also demonstrated that a malfolded substrate could still be

translocated. The latter result suggests that if proofreading of protein structure is an

important facet of ApH-dependent pathway export, it is substrate specific.

The versatility of the ApH-dependent pathway was further demonstrated during

the exploration of translocation of the integral thylakoid membrane protein Pftf plastidd

fusion/protein translocation factor). The precursor to Pftf, pPftf, carries an RRXFLK

motif typical of bacterial Tat pathway proteins. The integration of Pftf was shown by

several criteria to be ApH-dependent pathway specific, raising the question of how a

translocation mechanism that can export folded proteins can also recognize and integrate

a transmembrane domain. More surprising was the discovery that integration is not

dependent on the presequence but can be accomplished by N-tail export of the mature









protein. Integration of all forms of Pftf is inhibited by antibodies to Hcfl06, Tha4, and

TatC, but not by antibodies to cpSecY or cpOxalp, implying that all known components

of the ApH-dependent pathway are involved in its translocation (Summer et al., 2000).

Attempts to dissect translocation have led to a model in which at least two steps

are required for export. In the first step, which is not dependent on the trans-thylakoid

pH gradient, the precursor is bound to the cis side of the membrane (Ma and Cline, 2000;

Musser and Theg, 2000a). Subsequently upon establishment of the ApH, the precursor is

moved through the membrane and processed to mature size. Ma and Cline (2000) found

that the rate of transport of freshly added precursor is comparable to the rate of transport

from the bound state, making it unlikely that binding involves insertion into a putative

translocon. Biphasic kinetics of transport implies a slow step such as two-dimensional

diffusion within the membrane prior to the energy requiring transport step(s) (Musser and

Theg, 2000a). A solvent isotope effect implicates proton transfer as the rate-limiting step

in protein translocation (Musser and Theg, 2000b). When thylakoids are pre-incubated

with antibodies to either Hcfl06 (Ma and Cline, 2000) or cpTatC (Cline and Mori,

unpublished) the binding and subsequent transport (chase) of ApH-dependent pathway

directed precursors is inhibited. Pre-incubation of membranes with antibody to Tha4

does not inhibit binding, but does inhibit chase from the bound state. One interpretation

of these data is that precursor binds to a complex of Hcfl06 and TatC prior to interaction

with Tha4. However, the results are not definitive since binding of antibody to pathway

components could have multiple and complex effects.

Given the results from the studies of Musser and Theg (2000a) and Ma and Cline

(2000), it is now possible to put a slightly different interpretation on the results of an









earlier set of experiments by Berghofer and Klosgen (1999). When translocation

characteristics of a chimeric protein comprising the presequence of OE16 (also called

OE17) and the mature domain of OE23 (16/23) were investigated, two presumed

translocation intermediates were distinguished. Whereas the mature form of 16/23 was

expected to be a 23-kDa lumenal protein, in organello transport resulted primarily in a

27-kDa membrane associated protein. Thermolysin treatment of recovered thylakoids

yielded a 26-kDa degradation product. It was concluded that the stromal targeting

domain had been removed and export of the remainder of the protein had been

incomplete, leaving a few N-terminal amino acids on the cis side of the membrane.

When 16/23 was incubated on ice with isolated thylakoids, it associated tightly with the

membranes. The association also took place in the presence of nigericin. Thermolysin

treatment of reisolated membranes yielded a 14-kDa degradation product. The

degradation product was assumed to be inside the lumen. However, no controls were

reported to substantiate that assumption in spite of the fact that OE23 has been

demonstrated to yield a 22-kDa protease resistant fragment (Creighton et al., 1995). The

authors concluded that the 14-kDa fragment represented an early step in translocation not

dependent on functional translocase. Thus the precursor 16/23 may have associated with

the membranes in a pathway specific, but ApH independent manner as has since been

reported for a variety of constructs by Ma and Cline (2000). Upon establishment of

conditions favorable to protein translocation (250C and actinic light), export occurred.

Membranes recovered from a time course and treated with thermolysin yielded

progressively less 14-kDa degradation product and more 26-kDa degradation product.

Membranes were recovered following incubation of 16/23 with isolated thylakoids and









solubilized with digitonin. Samples prepared from those membranes were analyzed by

blue native polyacrylamide gel electrophoresis and visualized by radiography. The

radioactivity accumulated at two positions. Comparison of the location of those bands to

the location of known chloroplast protein complexes gave estimates of 560-kDa and 620-

kDa complexes. The authors assumed that the radioactivity was the result of the partially

translocated 16/23 previously demonstrated to give rise to the 26-kDa lumenal fragment.

However, the bands may have been the result of externally bound precursor that was

associated with ApH-dependent pathway components.


Summary and Perspective

Recent biochemical studies of the thylakoid ApH-dependent pathway and the

genetically related bacterial Tat pathway have emphasized the dichotomic relationship of

these translocation systems to previously known export pathways. The ApH-dependent

pathway, like the Sec system, is able to translocate both soluble and transmembrane

proteins, but unlike the Sec system, the ApH-dependent pathway and Tat pathways are

able to translocate folded proteins. While employing an amino terminal signal sequence

reminiscent of the eukaryotic, bacterial, and chloroplast Sec system signal peptides, the

ApH-dependent pathway requires no soluble accessory proteins and no NTP's. The

stimulatory effect of the proton motive force is variable in other export systems, but for

the ApH-dependent pathway, it is the sole energy source. The unique character of the

ApH-dependent pathway has led to much speculation about its mechanism of energy

utilization, the nature of its channel (if any), and its regulation. By investigating the

mode of initiation of export and by achieving a translocation intermediate, I have begun

to address the mechanism of substrate movement through the translocon. With the






18


integration of in vitro-translated radiolabeled ApH-dependent pathway components into

functioning membranes, I have created a tool with which component and substrate

interactions may be investigated.














CHAPTER 2
EVIDENCE FOR A LOOP MECHANISM OF PROTEIN TRANSPORT BY THE
THYLAKOID ApH-DEPENDENT PATHWAY




Introduction

Multiple models have been proposed to account for the apparent ability of the

ApH-dependent pathway to translocate folded proteins without compromising the trans-

thylakoid pH gradient. Upon initial cloning ofHcfl06, Settles et al. (1997) suggested

that the structure and topology of the protein made it a candidate for a receptor function.

Considering the bacterial Tat pathway, Berks et al. (2000b) described what they termed a

'sea anemone' model in which multiple copies of TatA/B/E form the pore with their C-

terminal regions gating the cytoplasmic side of the bilayer. An iris mechanism would

control the size of pore as individual subunits moved against each other. However, they

went on to cast doubt on their own model pointing out that a dynamic seal might be

insufficient to maintain the proton motive force. Musser and Theg (2000a) envisioned a

gated pore in the thylakoid. In their model the precursor is transferred to the pore cavity

and the pore sealed from the stromal side prior to its opening to the lumenal side. Berks

et al. (2000a) agreed with the idea of two proton-impermeable gates and proposed a more

detailed model for export via the bacterial Tat pathway. They suggested TatC as the site

of signal peptide recognition and TatB as "physically connecting, and thus mediating

communication between, the TatA and TatC subunits" (Berks et al., 2000a, p. 328).

Long-Fei Wu et al. (2000) favored TatA, TatB, or TatE as both receptors and gates with









TatC as the core structure of the channel. They suggested that the integration of the

transmembrane segments of the other Tat components might increase the size of the pore

in response to the translocating substrate.

Any model of the ApH-dependent pathway mechanism must not only explain

translocation of folded proteins, it must also explain integration of Pftf (Summer et al.,

2000). Clearly, the translocon does more than sense the presence of a precursor and

adjust a channel to accommodate it. The apparatus in some way monitors the substrate as

it initiates and as it continues translocation. The bacterial Tat pathway is able to

distinguish substrates that are properly assembled with cofactors or partner subunits

(Berks et al., 2000b). And the mechanism must be able to properly position an early

intermediate in export such that its signal peptide cleavage site is in position to be cut by

the lumenal (or periplasmic) processing protease.

Export-type systems, i.e. the Sec system and the endoplasmic reticulum (ER)

transport system, initiate transport via a transmembrane loop consisting of the signal

peptide and carboxyl-flanking region (Kuhn et al., 1994; Shaw et al., 1988). In contrast,

current models of protein import into mitochondria and chloroplasts depict insertion and

transport as proceeding amino terminus first (see Schatz, 1996 and Kouranov et al., 1996

for review); peroxisomal and endocytic processes are capable of transporting folded

proteins and even gold particles (Gietl, 1996). Here I present evidence that during

transport on the ApH-dependent pathway, the amino terminus of the precursor protein

remains on the cis side of the membrane. This is consistent only with the loop

mechanism used by export-type systems. Thus, my data suggest that the ApH-dependent









pathway shares a common evolutionary origin with other export pathways such as the

Sec systems and spontaneous insertion mechanisms.


Materials and Methods

Materials

All reagents, enzymes, and standards were from commercial sources. In vitro

transcription plasmids for precursors to pOE231 from pea, pPC from Arabidopsis, and

pOE17 from maize have been described elsewhere (Cline et al., 1993). Escherichia coli-

produced iOE23 has been previously described (Cline et al., 1993). pOE23 was produced

in E. coli (Cline et al., 1993) and then used for antibody preparation in rabbits. Antibody

to pOE17, prepared against the fusion protein pOE17-maltose binding protein, was the

generous gift of Dr. Alice Barkan (University of Oregon). Primers used in PCR reactions

were manufactured by DNAgency (Malvern, PA).

Construction of Chimeric Precursors

Coding sequences for recombinant proteins were constructed by PCR-based

methods using the above plasmids as templates. Amplifications were performed with Pfu

polymerase (Stratagene, La Jolla, CA). Cloned constructs were verified by DNA

sequencing. Sequencing was done with ABI Prism Dye Terminator cycle sequencing

protocols developed by Applied Biosystems (Perkin-Elmer Corp., Foster City, CA) and

an Applied Biosystems model 373 Stretch DNA Sequencer (Perkin-Elmer Corp.).

Sequencing of all clones on both strands was performed by the University of Florida






1 p-, i-, and m- are abbreviated designations for precursor-, intermediate precursor-, and
mature forms of the thylakoid proteins described in the text, respectively.









Interdisciplinary Center for Biotechnology Research (ICBR) DNA Sequencing Core

Facility.

The chimeric precursor m23p17 is a fusion of coding sequences for amino acids

constituting mOE23 (MAYGEAAN...TASSFSVA), a nine amino acid spacer

(QKEKNLGAE), and the complete pOE17 sequence. In an effort to eliminate internal

initiation in m23pl7, the methionines in the presequence of pOE17 were replaced. The

natural presequence beginning MAQAMASM.... was changed to LAQALASL.... Two

stages of PCR reactions were used in the construction of the clone for m23pl7. In the

first stage, two PCR reactions produced fragments coding for mOE23 (containing an

Xbal site near the 5' end) and for the nine amino acid spacer plus pOE17 (containing a

HindllI site near the 3' end). The forward primer for the nine amino acid spacer plus

pOE17 contained an overlap corresponding to the last 17 bases of mOE23. The two PCR

products were purified and spliced by overlap extension (SOE) in a third PCR reaction

(Horton et al., 1989). The SOE product was restricted and ligated in the SP6 direction

into pGEM 4Z cut with Xbal and HindllI.

The chimeric protein m23p is a fusion of coding sequences for the amino acids

constituting mOE23, the same nine amino acid spacer, and the complete presequence of

pOE17 (MAQAMASM...ALSQAARA). m23p was amplified in a single round of PCR

using as template a version of m23pl7 in which the original methionines were

unmodified.

Preparation of Radiolabeled Precursors by In Vitro Translation

Capped RNA for the various precursors was produced in vitro with SP6

polymerase essentially as described by Cline (1988). Translation in the presence of [3H]

leucine was in rabbit reticulocyte (Promega) or a wheat germ system (Cline et al., 1993),









as indicated in the figure legends. Translations were terminated by transfer to ice,

dilution threefold, and adjustment to import buffer [50mM HEPES (pH 8), 0.33 M

sorbitol] and 30 mM unlabeled leucine.

Preparation of Chloroplasts, Thylakoids, and Lysate

Intact chloroplasts were isolated from 9- to 10-day old pea seedlings (Laxton's

Progress 9) as described (Cline et al., 1993) and were resuspended in import buffer (IB).

Lysate and washed thylakoids were prepared from isolated chloroplasts (Cline et al.,

1993). Chlorophyll concentrations were determined according to Amon (1949).

Chloroplast Import and Thylakoid Protein Transport Assays

Import of radiolabeled precursors into isolated chloroplasts or transport into

washed thylakoids or chloroplast lysate was conducted in microcentrifuge tubes in a 25C

water bath illuminated with 70 [tE m-2s-1 white light (Cline et al., 1993) for 10 minutes or

the time indicated in the figure legend. For assays conducted in the presence of

inhibitors, chloroplast lysates were preincubated with azide (8 mM final concentration) or

nigericin (0.5 [tM final concentration) and valinomycin (1 LM final concentration) on ice

for 10 minutes prior to the addition of Mg-ATP (5 mM final concentration) and

radiolabeled precursor. For assays conducted in the absence of ATP, lysate (50 itg

chlorophyll in 50 [tl) and diluted translation product (25 tl) were each preincubated

separately for 10 minutes on ice with 1 U apyrase. Competition assays for thylakoid

transport were conducted as described previously (Cline et al., 1993). Assays were

generally terminated by transfer to 0C. Where indicated, recovered chloroplasts or

thylakoids were protease post-treated with thermolysin. Chloroplasts were repurified on

Percoll cushions; thylakoids were recovered by centrifugation.









Immunoprecipitation

Samples for immunoprecipitation were boiled one minute in SDS (final

concentration 1%) and then diluted 1:11 with 10 mM Tris/HCl (pH 7.5), 140 mM NaC1, 1

mM EDTA, 1% Triton-X100 and 1 mM PMSF. Rabbit antiserum (5 tl) was added, the

samples incubated over-night at 40C, and then rotated slowly for two hours at room

temperature. Protein A-Sepharose slurry (30 [l) in 10 mM Hepes/KOH (pH 8.0), 10 mM

MgCl2 was added and rotation continued at room temperature for one hour. The pellet

was recovered and washed four times with 10 mM Tris/HCl (pH 7.5), 150 mM NaC1, 2

mM EDTA and 0.2% Triton-X100 followed by washing twice with 10mM Tris/HCl (pH

7.5). The pellet was dissolved in SDS sample buffer and heated two minutes at 1000C.

The recovered supernatant was analyzed by SDS-PAGE/fluorography.

Analysis of Samples

Samples were subjected to SDS-PAGE and visualized by fluorography (Cline,

1986). Quantification of the radiolabeled protein in a gel band was accomplished by

scintillation counting of the excised and extracted gel band (Cline, 1986). In the time

course experiment, the relative molar quantity of m23p17 was calculated as the cpm

(minus background) in the m23p17 band divided by the number of leucines in m23p17.

The relative molar amount of each product was calculated as the number of cpms (minus

background) in each product band divided by the number of leucines expected in the

protein at that location. The relative molar percent of each protein is the relative molar

amount multiplied by 100 and divided by the relative molar amount of m23p17 at time

zero.









Results

A Chimeric Precursor Protein for Examination of Transport Topology Was
Constructed

To investigate the topology of transport on the ApH-dependent pathway, we constructed a

protein, m23p17, consisting of the mature domain of OE23 (mOE23) fused to pOE17

(Fig. 2-1). If the thylakoid Delta pH translocation machinery transports proteins via a

loop structure, the internally located LTD must engage the translocon while both mOE23

and mOE17 are on the cis side of the membrane. Subsequently the amino acids carboxy-

proximal to the LTD would enter the lumen. Cleavage by the lumenal processing

protease would release mOE17 on the trans side of the membrane, leaving the LTD and

amino acids amino-proximal to it, m23p, on the cis side (Fig. 2-1).

Incubation of m23p17 with isolated thylakoid membranes under transport-

permissive conditions produced two polypeptide products (Fig. 2-2A). At time zero only

the translation product, m23p 17, and a faint band apparently resulting from internal

initiation were visible. By 1 minute, polypeptides with the expected molecular weights

of m23p and mOE17 began to accumulate. Quantification of radioactivity in the bands

demonstrated that the two products accumulated progressively and stoichiometrically

throughout the 15-minute experiment (Fig. 2-2B). Furthermore, the relative molar

decrease in m23p17 was equal to the relative molar increase in the two products. This

processing was very efficient; approximately 70% of the substrate was cleaved to the two

products in the 15-minute assay. The identity of the products was confirmed by co-

migration with authentic polypeptides and by immunoprecipitation (Fig. 2-2C). The

presumptive m23p migrated identically to an authentic m23p translation product prepared

by recombinant methods (lanes 3 and 4) and was immunoprecipitated by antibodies to







A loop mechanism for transport.


OE23

N ^


OE17

STD LTD q -C


/


N


Figure 2-1. Model for translocation of a fusion protein, m23p17, by a loop initiation
mechanism. The mature form of OE23 was attached amino terminal to the full sequence
of pOE17. The model depicts the substrate and products as well as their relative
locations if the ApH-dependent pathway initiates transport with a loop.

OE23 (lane 8), but not by preimmune serum (lane 7) or by antibodies to OE17 (lane 9).
The mOE17 band migrated identically to authentic mOE17 (see below) and was
immunoprecipitated only by antibodies to OE17 (lane 9). The m23p17 present in
transport mixtures was immunoprecipitated by both OE23 and OE17 antibodies as


expected (lanes 8, 9).


- %r%


~












A

Time(mln) 0 1 15 2 25 3 5 10 15
TP
m23p17- q M

-m23p


- -mOE17


0 2 4 6 8 10 12 14
time (mm)


C

Mr
(kDa)
66-
45-
36-
29- mW -


- i i


lanes 1 2 3 4 5 6


7 8 9


Figure 2-2. Temporal appearance and identities of peptides produced during a
transport reaction with m23p17. (A) Rabbit reticulocyte-produced m23p17 was
incubated with washed thylakoid membranes in an illuminated bath at 250C. Samples
were taken at the times indicated, boiled in SDS buffer, and analyzed by SDS-PAGE and
fluorography. The radiolabeled precursor (TP) represents an amount equivalent to that in
each assay sample. (B) Quantification of bands shown in (A). (C) m23p17 and m23p
were generated by translation with rabbit reticulocyte and are seen in lanes 1 and 4,
respectively. A 10-min transport assay with m23p17 was divided into 5 aliquots.
Samples for lanes 2 and 3 were boiled in SDS. The remaining material was processed for
immunoprecipitation (see Materials and Methods). Lane 3 contains a mixture of
transport products (90% of the amount in lane 2) and translation-generated m23p (equal
to the amount in lane 4). Lane 5 and 6 contain m23p subjected to immunoprecipitation
with antibodies to OE23 and OE17, respectively. Lanes 7, 8 and 9 contain transport
assay samples subjected to immunoprecipitation with preimmune, anti-OE23, and anti-
OE17, respectively.




In previous studies with ER membranes, the signal peptide was further cleaved in


the hydrophobic core prior to release of the amino terminal portion into the supernatant


p -m23p17

w -m23p


-- -mOE17









(Lyko et al., 1995). No processing of the LTD or degradation of m23p was apparent in

our experiments. We estimate that our gel analysis would have detected the loss of as

few as six amino acids.

m23p Is Localized to the Cis Side of the Membrane, whereas mOE17 Is in the
Lumen

To determine the location of the above products, soluble and membrane fractions

were obtained from a transport assay. In time course reactions such as the one shown in

Fig. 2-2, all of the mOE17 and most of the m23p were recovered with thylakoids (not

shown). Accordingly, thylakoids from a 10-minute transport assay were treated with

thermolysin to distinguish surface-exposed from lumenal species (Fig. 2-3). The m23p17

and m23p bands were degraded by protease treatment, indicating that they were exposed

to the stromal surface of the membranes, whereas the mOE17 band was protease

resistant, indicating a lumenal location (lane 5). When the thylakoid lumen was opened

by sonication or 1% Triton X-100, the mOE17 was degraded by protease (data not

shown). For comparison, in vitro translated pOE17 (lane 6) was similarly assayed with

thylakoids (lanes 9 and 10). Membrane-associated pOE17 and mOE17 displayed the

expected protease sensitivity and insensitivity, respectively.

Also shown in Fig. 2-3, m23p17 was not imported into isolated chloroplasts and

even failed to bind significantly to the chloroplast surface (lanes 2 and 3). In chloroplast

import experiments with rabbit reticulocyte translated m23pl7, a very faint band at the

location of mOE17 was produced (data not shown). This apparently resulted from import

of a small amount of internally initiated pOE17 that occurs in the reticulocyte system (see

Fig. 2-2). These results can be compared to the robust import achieved with the authentic

pOE17 precursor (lanes 7 and 8). The lack of import or binding of m23p17 by










chloroplasts is apparently due to their inability to recognize the internally located transit

peptide. This agrees with current models of organellar import in which translocation is

initiated at the N terminus of the targeting peptide, followed by movement of the

precursor linearly through the pore (Kouranov et al., 1996).


m23p17


TP Chloroplasts Thylakoids

V.o.,..


pOE17


TP Chloroplasts Thylakoids


-m23p17



-m23p
-pOE17


lanes: 1 2
Protease: -


4 5
+


6 7 IIIII

6 7 8
+


4 0 -mnOE17

9 10
+


Figure 2-3. Transport of m23p17 delivered mOE17 to the lumen, leaving m23p on
the stromal face of the thylakoids. Wheat germ-translated m23p17 and pOE17 were
assayed for import into isolated chloroplasts and transport into isolated thylakoids at 250C
for 10 minutes (see Materials and Methods). Recovered chloroplasts or thylakoids were
treated with or without thermolysin as shown below the panel. Lanes 1 and 6 represent
1% of the m23p17 and pOE17, respectively, added to each assay. Lanes 2 through 5 and
7 through 10 contain chloroplasts or membranes equivalent to 6% of that present in each
assay.



Production of m23p and mOE17 from m23p17 Results from ApH-dependent
Pathway Transport.

Because of the alterations necessary to make m23p17, it was important to

determine if m23p and mOE17 were produced as a result of transport via the ApH-

dependent pathway. Assays were conducted under conditions diagnostic for











A light: + + + +
ATP: + + + + [iOE23] ipM
nig/val: +
SE: + + + + + TP 0 U .2 .4 .8 2
apyrase: + + m23p17- -
azide: +
TP
m23pl7- -~ -.- -2p

S -m23p ....... -E17


..- - -mOE17
pOE17- ....... -mOE23

- -mOE17
pPC- --. -. -. TP 0 U 2
pPC- -
-mPC
lanes: 1 2 3 4 5 6 7 -mPC



Figure 2-4. m23p17 was transported exclusively on the ApH-dependent pathway.
(A) Transport of wheat germ-produced precursors across thylakoid membranes was
conducted for 10 minutes at 250C in light or darkness as shown, with lysate to provide
stromal extract (SE), or with twice-washed thylakoids (see Materials and Methods).
Assay conditions were designed to examine the requirement for SE, ATP, a ApH, or
sensitivity to azide. Apyrase was used to eliminate residual ATP in lysate and translation
products. These conditions are designated above the fluorogram panels. Recovered
thylakoids were washed and analyzed by SDS-PAGE and fluorography. The
radiolabeled precursor (TP) represents 1% of the amount in each assay. Lanes were
loaded with recovered thylakoids equivalent to 6.7% of each assay. (B) Transport
competition assays were conducted with chloroplast lysates and 5 mM ATP in the
presence of increasing concentrations of unlabeled iOE23 for 10 minutes at 250C in the
light. Thylakoid membranes were washed and analyzed by SDS-PAGE and
fluorography. Lanes were loaded with TP and sample amounts as in (A) above. The final
concentration of iOE23 competitor is indicated above the fluorograms. The lane
designated U represents a control assay containing no iOE23 competitor, but containing
urea (U) at 167 mM (equal to the concentration of urea in the 2 ptM competition assay).



ApH- mediated transport (Fig. 2-4). In Fig. 2-4A the energy and stromal requirements for

production of m23p and mOE17 are compared to those for transport of ApH-dependent

pathway-directed pOE17 and cpSec pathway-directed pPC. As with pOE17 transport,









production of m23p and mOE17 was abolished by ionophores that dissipate the trans-

thylakoid ApH (lane 4) and was unaffected by removal of ATP (lane 5), by the absence of

stromal extract (lane 6), or by sodium azide (lane 7), a SecA inhibitor (Oliver et al.,

1990). In contrast, cpSec-mediated pPC transport was only slightly diminished by

ionophores, but abolished by removal of ATP or stromal extract (the source of

approximately 90% of cpSecA) and diminished by addition of sodium azide.

In the experiment shown in Fig. 2-4A, only the recovered membranes were

analyzed. An unexpected result was that removal of stromal components and thorough

washing of thylakoids used for the assay, while without effect on the amount of mOE17

produced from m23p17 (top panel, lane 6), resulted in much less membrane-associated

m23p. In other experiments (not shown), we verified that m23p was indeed produced

under these conditions.

As an additional test of pathway utilization by m23pl7, precursor competition

studies were conducted. Precursor proteins utilizing the ApH-dependent pathway are able

to compete with one another for transport, but not with cpSecA utilizing precursors. As

shown in Fig. 2-4B, the ApH-dependent pathway intermediate iOE23 competed for

transport of pOE23 and m23pl7, but not pPC transport. Taken together, these results

demonstrate that transport of the mOE17 moiety of m23p17 occurs via the ApH-

dependent pathway. Moreover, they substantiate that production of the cis-localized

m23p is a transport-coupled phenomenon.


Discussion

A hallmark of a loop mechanism of insertion is that the amino-proximal peptide

flanking the signal sequence remains on the cis side of the membrane, while the carboxyl









flanking polypeptide is transported across the membrane. The loop mechanism was

demonstrated for the ER by expression in HeLa cells of a vesicular stomatitis virus

glycoprotein (VSV G) with a non-cleavable amino terminal extension (Shaw et al.,

1988). The VSV G domain was translocated to the ER lumen, whereas the amino

terminal extension remained exposed to the cytoplasmic side of the membrane. The

existence of an inserted loop in the ER was further demonstrated by Mothes et al. (1994)

who used photo-reactive crosslinkers to show that the polypeptide chain spanned the ER

membrane twice prior to, but only once following signal sequence cleavage. Kuhn et al.,

(1994) demonstrated a loop mechanism for protein translocation by the Sec pathway of E.

coli with a similar strategy to that employed in our studies; i.e. a chimeric precursor

protein having a large amino terminal extension derived from ribulokinase was fused to

proOmpA.

Similar to the results obtained with the ER system and the E. coli system, the

amino terminal domain of m23pl7 did not interfere with signal peptide recognition or

with translocation by the ApH-dependent pathway mechanism of thylakoids. In addition,

the m23p domain remained outside of the thylakoids, whereas the OE17 domain was

translocated into the lumen (see Fig. 2-1 for model). In studies with ER membranes, the

amino terminus remained transiently associated with the membranes prior to release into

the supernatant (Lyko et al., 1995). We also observed binding of transport-generated

m23p to membranes that varied with the specific conditions of the assay (Fig. 2-4A).

Membrane-bound m23p was tightly associated, as much of it remained bound even after

washing with high salt or urea. However the physiological relevance of such binding is

unclear because m23p produced directly by in vitro translation also bound tightly to the









membrane and was resistant to the same salt and urea extractions (Fincher and Cline,

unpublished results). Thus, although it is possible that transport-produced m23p remains

transiently associated with the translocon as depicted in Fig. 2-1, we cannot currently

distinguish this from non-specific association with the membrane surface.

The original loop model as proposed by Inouye and Halegoua (1980) describes a

loop formed across the bilayer with the hydrophobic region of the signal peptide

extending through the bilayer on one side and the first several amino acids of the mature

domain of the substrate protein extending through the bilayer on the opposite side as

depicted for m23p17 in Fig. 2-1. In light of evidence that the Tat pathway and ApH-

dependent pathway are able to translocate folded proteins, the availability of the amino

terminus of the mature protein to form the carboxyl side of the loop is open to question.

The extent of folding of OE17 in the m23p17 construct is unknown. It has been noted

that where crystal structures of bacterial Tat substrates are known, the mature protein

lacks a disordered N-terminus that could support loop formation, but that the structure of

the precursor might be somewhat different (Berks et al., 2000b). Clearly, the data I have

presented indicate that a loop was formed, although the precise amino acids involved in

loop formation and the location of the loop (whether in the membrane bilayer, in a

receptor site or at a protein-lipid interface) remain undetermined.

The unlooping model was put forward by de Kruijff and coworkers (de Vrije et

al., 1990) as an alternative hypothesis for protein translocation through Sec-type systems.

In their model, the positively charged N-domain interacts electrostatically with the

negatively charged phospholipids. That interaction is followed by insertion of the H-

domain as a loop reaching only into the hydrophobic core of the bilayer. The passage of









the peptide through the bilayer locally perturbs lipid organization. The H-domain

unloops pulling the mature region of the substrate into the hydrophilic channel (van

Voorst and de Kruijff, 2000). Since the alpha carbons of arginine residues are potentially

able to reside one or more helical turns within the membrane while the charged group

reaches the lipid headgroup region (Monne et al., 1998), the H-domain of m23p17 may

be effectively lengthened by several residues. The bending of the H-domain from one ca-

helix to effectively two (called helix-break-helix) is aided by amino acids common to

hydrophobic domains such as glycines and series (Izard and Kendall, 1994). Data from

the experiments I have described with m23p17 as well as data from the E. coli Sec

system (Kuhn et al., 1994) and the mammalian ER (Shaw et al., 1988; Mothes et al.,

1994) are reconcilable with the unlooping mechanism of protein entry. The conclusion

that remains is that in all three export pathways and in spontaneous translocation

(Thompson et al., 1998), transport is initiated by signal peptide mediated loop formation.














CHAPTER 3
THE CHARACTERIZATION OF A TRANSLOCATION INTERMEDIATE ON THE
ApH-DEPENDENT PATHWAY


Introduction

The biochemistry of the chloroplast ApH-dependent pathway has been studied by

defining its signal sequence specificity, energetic requirements, topology oftranslocation

initiation, and its ability to transport folded proteins. While its use of N-terminal

targeting sequences and initiation of translocation via a loop mechanism point to an

evolutionary relationship with export pathways such as that of the endoplasmic reticulum

(ER) and bacterial Sec system, other pathway characteristics set it apart. The ability of

the ApH-dependent pathway to utilize only the trans-thylakoid pH gradient as an energy

source is unique among known protein translocation systems. The recently elucidated

capacity to transport folded proteins (Clark and Theg, 1997; Hynds et al., 1998) is rare

among protein export systems, but has since been documented in the bacterial twin-

arginine translocation pathway to which the thylakoid pathway is evolutionarily related

(Berks et al., 2000a for review; Thomas et al., 2001).

The translocation of folded proteins has been clearly demonstrated in peroxisomal

import pathways (Gietl, 1996, for review). In spite of the apparent ability of peroxisomes

to translocate even large complexes, a translocation intermediate has been characterized

in the specialized peroxisomes called glyoxysomes (Pool et al., 1998). Observation of

the translocation intermediate was achieved using a fusion protein partially consisting of

the immunoglobulin G (IgG) binding domains of Staphylococcus aureus protein A. The









authors speculated that the ability to form such an intermediate was not dependent on the

tightly folded nature of protein A, but on the unique nature of the protein itself. Protein

A had previously been used to form a translocation intermediate and to purify

components of the chloroplast protein import machinery (Schnell and Blobel, 1993).

Schilke et al. (1997) constructed a chimeric precursor by linking the presequence of a

mitochondrial precursor via glutathione S-transferase (GST) to a polypeptide containing

multiple IgG binding domains of protein A. Import of the precursor into mitochondria

resulted in processing and formation of a translocation intermediate both in vitro and in

vivo. Formation of the translocation intermediate engendered "zippering" of the outer

and inner mitochondrial membranes.

Translocation intermediates are an important tool in the elucidation of transport

mechanisms and the isolation and identification of machinery components. I have

constructed a series of proteins consisting of pOE17 and the IgG binding domains of

protein A. The constructs differ in the nature and length of the linker region between the

two domains. Using these recombinant proteins, I have demonstrated the formation of

translocation intermediates on the ApH-dependent pathway.


Materials and Methods

Preparation of Chloroplasts, Thylakoids, and Lysate

Intact chloroplasts were isolated from 9- to 10-day old pea seedlings (Laxton's

Progress 9) as described (Cline et al., 1993) and were resuspended in import buffer.

Lysate and washed thylakoids were prepared from isolated chloroplasts (Cline et al.,

1993). Chlorophyll concentrations were determined according to Amon (1949).









Construction of Chimeric Precursors

Cloning and analysis of DNA products were by standard molecular biology

procedures (Sambrook et al., 1989). Amplifications were performed with Pfu polymerase

(Stratagene, La Jolla, CA). Cloned constructs were verified by DNA sequencing.

Sequencing was done with ABI Prism Dye Terminator cycle sequencing protocols

developed by Applied Biosystems (Perkin-Elmer Corp., Foster City, CA) and an Applied

Biosystems model 373 Stretch DNA Sequencer (Perkin-Elmer Corp.). Sequencing of all

clones on both strands was performed by the University of Florida Interdisciplinary

Center for Biotechnology Research (ICBR) DNA Sequencing Core Facility.

The cloning of protein A-containing constructs was a multi-stage process using

several templates. In the first step, maize pOE17 (Cline et al., 1993) and mouse

dihydrofolate reductase (DHFR) were used as templates in an SOE (Horton et al., 1989)

reaction1. The final product, pl7-DHFR, was ligated in the SP6 direction into pGEM 4Z.

The template pl7-DHFR was amplified by PCR with a 5' primer having an engineered

EcoRI restriction site and a 3' primer with coding for Kodak FLAG epitope (amino acid

sequence of FLAG epitope: DYKDDDDK). The 3' primer overlapped DHFR so that the

final product contained all of pOE17 and 22 amino acids of DHFR followed by the

FLAG epitope. The resultant DNA, pl7-FLAG, was purified, restricted with EcoRI and

Hincl, and ligated in the SP6 direction into pGEM 4Z that had been cut with EcoRI and

Hinci.

I made pl7A-protA by PCR and Sphl-mediated splicing using maize pOE17 in

pGEM 4Z and pS/protA in pET21b (Schnell and Blobel, 1993) as templates. I amplified


1 pl7-DHFR was cloned by Michael McCaffery in 1993.









pOE17 using the same 5' primer as employed in production of pl7-FLAG with an

engineered EcoRI restriction site and a 3' primer having an engineered SphI restriction

site. The resultant intermediate contained all but the final eight amino acids of pOE17. I

amplified pS/protA using a 5' primer having an engineered SphI restriction site and a 3'

primer having an engineered BamHI restriction site. I cut the intermediates with SphI

and ligated. From the ligation products, I selected the band of appropriate size (1659

base pairs), purified the DNA, restricted it with EcoRI and BamHI, and ligated it in the

SP6 direction into pGEM 4Z that had been cut with EcoRI and BamHI.

The construct pl7-L-protA was cloned in multiple steps including SOE and Sphl-

mediated splicing using a pBluescript SK phagemid (Promega) intermediate. The

pBluescript SK phagemid was used as an intermediate cloning vector because of the

presence of an SphI restriction site in pGEM 4Z that would have complicated cloning.

The template pl7-FLAG was amplified by PCR using a 5' primer having EcoRI and Ndel

restriction sites and a 3' primer with sequence that amplified coding for all of pOE17 and

19 amino acids from DHFR (but none of FLAG). The 3' primer included an overlap for

a portion of the 5' primer for pGEX-2TK (Pharmacia). A portion of the pGEX-2TK

template that encodes 25 amino acids including the thrombin site was amplified using a

5' primer with the stated overlap and a 3' primer having an SphI restriction site. The two

PCR products were purified and spliced by SOE in a third PCR reaction to yield pl7-L.

The construct described in the previous paragraph, pl7A-protA, was cut from pGEM 4Z

using EcoRI and BamHI, then ligated into pBluescript SK phagemid that had been cut

with EcoRI and BamHI. The product was restricted with EcoRI and SphI, resulting in a

linear plasmid having an open SphI site followed by pS/ProtA. I then cut pl7-L with









EcoRI and SphI and ligated it into the linear pBluescript SK phagemid. The result was

pl7-L-protA in pBluescript SK phagemid. I cut pl7-L-protA from pBluescript SK

phagemid with EcoRI and BamHI and ligated it in the SP6 direction into pGEM 4Z that

had been cut with EcoRI and BamHI.

The construct pl7-LF-protA was cloned similarly to pl7-L-protA. The template

pl7-FLAG was amplified using a 5' primer having EcoRI and Ndel restriction sites and a

3' primer with sequence that amplified coding for all of pl7-FLAG. The 3' primer

included an overlap for a portion of the 5' primer for pGEX-2TK. A portion of the

pGEX-2TK template that encodes 28 amino acids including the thrombin site was

amplified using a 5' primer with the stated overlap and a 3' primer having an SphI

restriction site. The two PCR products were purified and spliced by SOE in a third PCR

reaction to yield pl7-LF. The construct pl7A-protA was cut from pGEM 4Z using EcoRI

and BamHI, and then ligated into pBluescript SK phagemid. The product was cut with

EcoRI and SphI, resulting in a linear plasmid having an open SphI restriction site

followed by pS/ProtA. I subsequently restricted pl7-LF with EcoRI and SphI and ligated

it into the linear pBluescript SK phagemid. The result was pl7-LF-protA in pBluescript

SK phagemid. Sequencing revealed an error in the clone in the pOE17 portion near the

5' end. I cut the DNA just 3' to the error with SaclI and at its 3' end with BamHI to yield

a 1704 base pair product. I restricted pl7-L-protA in pGEM 4Z with SaclI and BamHI

resulting in a linear plasmid with the 5' end of pOE17 up to the SacII restriction site. I

ligated the 1704 base pair insert into the pGEM 4Z plasmid to get pl7-LF-protA in the

SP6 direction in pGEM 4Z.









Truncations of pl7-LF-protA termed LF 1, LF2, and LF3 were cloned by PCR

using pl7-LF-protA in pGEM 4Z as the template.2 The 5' primer was made to the

sequence of pGEM 4Z upstream of the clone start site. Because protein A has multiple

domains of similar sequence, it was possible to make all three clones using a single 3'

primer. Products of the PCR having sizes 1197 base pairs, 1380 base pairs, and 1554

base pairs were purified, restricted with EcoRI and BamHI, and ligated in the SP6

direction into pGEM 4Z that had been restricted with EcoRI and BamHI to yield LF1,

LF2, and LF3, respectively.

The truncation of pl7-LF-protA named LFO includes the complete linker region

of the full construct without any of protein A3. It was amplified from LF 1 using a 5'

primer that initiated in the pGEM-4Z vector and a 3' primer that initiated within the

linker region and contained an engineered BamHI site. The PCR product was purified,

restricted with EcoRI and BamHI, and ligated in the SP6 direction into pGEM 4Z that had

been restricted with EcoRI and BamHI to yield LFO.

The construct pLF(V/R) was cloned by PCR using SOE4. The template was pl7-

LF-protA; the 5' and 3' primers were to vector regions upstream and downstream,

respectively, of the clone. Internal primers were designed to change the single amino

acid, valine 229, to arginine and yield products with overlapping sequence in the first set

of PCR amplifications. The second round of PCR yielded pLF(V/R). In a similar set of

reactions, alanine 290 was mutated to arginine yielding pLF(A/R). The PCR products

were purified, restricted with EcoRI and BamHI, and ligated in the SP6 direction into



2 LF1, LF2, and LF3 were cloned by Justin Delille and Mike McCaffery in 1999.
3 LFO was cloned by Mike McCaffery in 1999.
4 pLF(V/R), pLF(A/R), and pLF(V/R)(A/R) were cloned by Mike McCaffery in 2000.









pGEM 4Z that had been restricted with EcoRI and BamHI. The dual change

pLF(V/R)(A/R) was constructed by SphI digestion and splicing of the single change

constructs.

Clone pLF(V/R)(A/R) was used as template for the cloning of pl7tp-protA by

SOE.5 A 5' primer having an engineered SstI site and a 3' primer including an overlap

region for the SOE were used to amplify nucleotides encoding the transit peptide of

pOE17. A 5' primer with the stated overlap for the SOE was used to amplify the protein

A moiety beginning with glutamine 308 from pLF(V/R)(A/R). The 3' primer used for

the amplification of the protein A moiety bound in the vector. After the SOE reaction,

the product was purified, cut with SstI and XbaI, and ligated in the SP6 direction into

pGEM 4Z that had been cut with SstI and XbaI.

Preparation of Radiolabeled Precursors

Coupled transcription/translation with wheat germ TnT (Promega) or rabbit

reticulocyte TnT (Promega) in the presence of3H leucine or, where specified, 35S

methionine (NEN Life Science Products) was performed following the manufacture's

guidelines. In vitro transcription with SP6 RNA polymerase (Promega) and translation

with wheat germ lysate (Promega) in the presence of 3H leucine was performed following

the manufacture's guidelines. Translation products were diluted with one volume 60 mM

leucine (or 60 mM methionine for 35S labeled proteins) in 2X import buffer (IX = 50 mM

HEPES, KOH, pH 8.0, 0.33 M sorbitol) prior to use.


5 pl7tp-protA was cloned by Mike McCaffery in 2001.











Chloroplast Import and Thylakoid Protein Transport Assays

Import of radiolabeled precursors into isolated chloroplasts or transport into

washed thylakoids or chloroplast lysate was conducted in microcentrifuge tubes in a 25C

water bath illuminated with 70 ptE m-2s-1 white light (Cline et al., 1993) for 10 min or the

time indicated in the figure legend. For assays conducted in the presence of inhibitors,

chloroplast lysates were preincubated with nigericin (0.5 [tM final concentration) and

valinomycin (1 LM final concentration) on ice for 10 min prior to the addition of Mg-

ATP (5 mM final concentration) and radiolabeled precursor. Competition assays for

chloroplast import were conducted as described previously (Cline et al., 1993). Assays

were generally terminated by transfer to an ice bath. Termination with HgC12, where

indicated in the figure legend, was used when it was necessary to rapidly halt both

transport and all intra-lumenal reactions (Reed et al., 1990). Where indicated, recovered

chloroplasts or thylakoids were protease post-treated with thermolysin. Chloroplasts

were repurified on Percoll cushions; thylakoids were recovered by centrifugation.

Immunoprecipitation

Samples for immunoprecipitation were heated at 800C for 5 min in SDS (final

concentration 1%) and then diluted 1:11 with 10 mM Tris/HCl (pH 7.5), 140 mM NaC1, 1

mM EDTA, 1% Triton-X100 and 1 mM PMSF. Rabbit antiserum (5 [tl) was added and

the samples were rotated slowly over-night at 40C. Protein A-Sepharose slurry (30 [[l) in

10 mM Hepes/KOH (pH 8.0), 10 mM MgCl2 was added and rotation continued at 40C

for two hours. The pellet was recovered and washed four times with 10 mM Tris/HCl

(pH 7.5), 150 mM NaC1, 2 mM EDTA and 0.2% Triton-X100 followed by washing twice









with 10mM Tris/HCl (pH 7.5). The pellet was dissolved in SDS sample buffer and

heated 2 min at 1000C. The recovered supernatant analyzed by SDS-PAGE and

fluorography.


Results

Fusion Proteins Consisting of pOE17 and Protein A Were Constructed Having
Varying Linkers.

The seventeen kilodalton component of the oxygen evolving complex, OE17 is encoded

in the nucleus, synthesized in the cytoplasm in its precursor form, pOE17, and imported

into chloroplasts. Inside the chloroplasts, the ApH-dependent pathway exports it to the

thylakoid lumen. I created several fusion proteins using maize pOE17 and S. aureus

protein A. When referred to collectively, these constructs are termed "17-protA";

however, each individual fusion protein is described and given a specific label. The

OE17 domain of pl7A-protA has eight amino acids truncated from its carboxyl end. All

other chimeras described include the entire coding sequence of pOE17. The overall

design of the constructs is pictured in Fig. 3-1A. Each chimera differs in the length and

content of its linker region as described in Fig. 3-1B. The complete amino acid sequence

of pl7-LF-protA is detailed in Fig. 3-1C with annotations.

S. aureus protein A has five IgG binding domains (Moks et al., 1986). The full-

length chimeric constructs described in Fig. 3-1A each comprise four complete binding

domains plus a portion of the fifth domain. Each IgG binding domain contains three

alpha helices (Stahl and Uhlen, 1997, Alonso and Daggett, 2000) indicated by color in

Fig. 3-1C. There is very high amino acid homology among the domains as can be

discerned by inspection. Truncated versions of pl7-LF-protA were created having no

protein A IgG binding domains (LFO), one IgG binding domain (LF 1), two IgG binding













Overall Design of Constructs


1 69


210 242


presequence OE17A linker


1 69


Protein A


218 292


p17-L-protA





p17-LF-protA


presequence OE17 linker Protein A


1 69 218 308 602
I I I I


presequence OE17


linker


Protein A


Details of Linker Regions

p17A-protA
210 211 AC created by Sphl splicing
212 241 SQVLKELDEWAAYPQAFVRIIGFDNVRQV from small subunit RUBISCO


p17-L-protA
218-236 GMVRPLNCIVAVSQNMGIG from Dihydrofolate Reductase
237 261 WQATFGGGDHPPKSDLVPRGSRRAC from pGEX-2TK includes
thrombin site LVPRGS and modification of kinase site from RRASV
to RRAC by Sphl splicing
262- 291 SQVLKELDEWAAYPQAFVRIIGFDNVRQV from small subunit RUBISCO


p17-LF-protA
218 239 GMVR PLNCIVA(V)(R)SQNMGIG KNG from Dihydrofolate Reductase
240 249 DYKDDDDEED from Kodak FLAG DYKDDDDK with change
of K to E and two additional amino acids added by PCR mutagenesis
250 277 LQGWQATFGGGDHPPKSDLVPRGSRRAC from pGEX-2TK includes
thrombin site LVPRGS and modification of kinase site from RRASV
to RRAC by Sphl splicing
278 307 SQVLKELDE WA(A)(R)YPQAFVRIIGFDNVRQV from small subunit RUBISCO



Figure 3-1. Design of fusion proteins. (A) The carboxyl end of truncated (OE17A) or
full length pOE17 followed by a linker region was fused to a portion of the IgG binding
segment of S. aureus protein A. (B) The linkers in each fusion protein differed in length
and sequence. The orange amino acids are regions of relatively high hydrophobicity.
Within those regions, parenthetical amino acids in orange are those present in p17-LF-
protA which were changed to arginines in constructs LF(V/R), LF(A/R), and
LF(V/R)(A/R). Lavender amino acids are the site of thrombin cleavage.


p17A-protA











C Sequence of p17-LF-protA


MAQAMASMTGLSQGVLPSRRADSRTRTAWIVRA SAEGDAVAQAGRR AVIGLVATGIVGGAL SQAARA
stromal targeting domain N-domain H-domain C-domain

ETVKTIKIGAPPPPSGGLPGTLNSDQARDFDLPLKERFYQQPLPPAEAAARVKTSAQDIINLKPLIDKKAWPY
begin mature OE17

VQNDLRLRASYLRYDLKTVIASKPKEEKKSLKELTGKLFSTIDDLDHAAKIKSTPEAEKYFAATKDALGDVLAKIG

GMVRPLNCIVAVSQNMGIGKNGDYKDDDDEEDLQGWQATFGGGDHPPKSDLVPRGS
begin linker region
RRACSQVLKELDEWAAYPQAFVRIIGFDNVRQVQCIDSGGVTPAANA
end LFO
AQHDEAQQNAFYQVLNMPNLNADQRNGFIQSLKDDPSQSANVLGEAQKLNDSQAPKAD
begin mature protein A end LF1

AQQNNFNKDQQSAFYEILNMPNLNEAQRNGFIQSLKDDPSQSTNVLGEAKKLNESQAPKAD
end LF2

NNFNKEQQNAFYEILNMPNLNEEQRNGFIQSLKDDPSQSANLLSEAKKLNESQAPKAD
end LF3
NKFNKEQQNAFYEILHLPNLNEEQRNGFIQSLKDDPSQSANLLAEAKKLNDAQAPKAD

NKFNKEQQNAFYEILHLPNLTEEQRNGFIQSLKDDPGNSRGTMDLE



Figure 3-1. Design of fusion proteins (continued). (C) The full sequence of pl7-LF-
protA is detailed. The presequence includes the stromal targeting domain in gray and the
tripartite lumenal targeting domain with the N-domain in dark red, the H-domain in dark
blue, and the C-domain in brown. The linker region is colored as in (B). The first amino
acid from the protein A moiety, glutamine 308, is turquoise. Protein A comprises five
IgG binding domains E, D, A, B, and C named in order from amino to carboxyl terminus
(Stahl and Uhlen, 1997). Each of these domains contains three alpha helices show in
light red, light green and light blue. Only the first two helices of domain C are included
in this construct. The carboxyl ends of truncated proteins LFO, LF1, LF2, and LF3 are
indicated.



domains (LF2), and three IgG binding domains (LF3). The point of truncation for each

of these constructs is indicated in Fig. 3-1C.

Fractionation of Chloroplasts Following Import Demonstrated a Thylakoid
Localized Mature Form of 17-protA.

The panels of Fig. 3-2 display the results of assays conducted on three versions of

p17-protA. Import of in vitro-translated p17-protA into intact chloroplasts or transport









into washed thylakoids resulted in processing of the precursor to mature size.

Fractionation of chloroplasts revealed that the mature sized product was primarily

associated with the membrane (Fig. 3-2, lane 4), although some remained soluble in the

Fractionation of Import


p17A-protA -

ml 7A-protA -


pl7-L-protA -

ml7-L-protA -



pl7-LF-protA -

m17-LF-protA -


Protease


Lane


*-mow


- +


- +


1 2 3 4 5


6 7


Figure 3-2. Intact chloroplasts and washed thylakoids process pl7-protA resulting
in a membrane associated mature form. Rabbit reticulocyte TnT generated precursors
were incubated with intact chloroplasts (import) or washed thylakoids (transport) for 10
min. Import and transport assays were divided. Chloroplasts were recovered and
aliquots were subfractionated by centrifugation to yield stroma and membranes.
Membranes were protease treated with thermolysin where indicated below the panels.
Samples were analyzed by SDS-PAGE (7.5% acrylamide) and fluorography.









stromal fraction (Fig. 3-2, lane 3). Membrane associated ml7-protA was sensitive to

treatment with thermolysin (Fig. 3-2, lane 5 and lane 6). When ml7A-protA

(approximately 52 kDa) is associated with intact chloroplasts or with stroma, it runs in a

curved band on an SDS-PAGE gel as can be seen in Fig. 3-2, lanes 1-3. The band also

runs lower on the gel in the presence of stroma than in washed thylakoids (Fig. 3-2, lanes

1, 2, and 3 versus lane 4 and 7). This phenomenon is probably due to the proximity of

other stromal proteins, notably the large subunit of ribulose bisphophate

carboxylase/oxygenase (RUBISCO).

Processing of pl7-protA to the Mature Size Occurs Only under Transport
Permissive Conditions.

For ApH-dependent pathway precursors, export to the thylakoid lumen is

dependent on the trans-thylakoid pH gradient. Dissipation of that gradient by ionophores

such as nigericin results in the accumulation of the stromal intermediate form as

demonstrated in lanes 4 and 5 of Fig. 3-3 for ApH-dependent pathway precursor pOE23.

Import of pl7-protA in the presence ionophores resulted in the accumulation of

intermediate forms (lanes 4 and 5). (As seen in Fig. 3-2, the *ml7A-protA runs lower in

the presence of stromal proteins.) The accumulation of the intermediate in the stroma

implies that processing to the mature size is dependent on access to the thylakoid lumenal

processing protease. The inhibition of transport by ionophores was incomplete for OE23

and 17A-protA. As a result, some mOE23 can be seen in the lumen in lanes 4 and 6.

Because ml7A-protA is able to back out of the lumen, it is visible along with il7A-protA

in the chloroplasts and stroma of ionophore treated samples (lanes 4 and 5).

Precursor proteins utilizing the ApH-dependent pathway are able to compete with

one another for thylakoid transport, but not with cpSecA utilizing precursors. It has been









demonstrated that the import of pOE23 into intact chloroplasts can concentrate the

stromal intermediate for thylakoid transport to supersaturating levels (Cline et al., 1993).

A model of in organello competition is displayed in Fig. 3-4A. In the presence of

increasing concentrations of the unlabeled ApH-dependent pathway competitor pOE23,

the processing of labeled pOE17 and pl7-ProtA is progressively inhibited (Fig. 3-4A).

When 0.2 to 0.4 tlM pOE23 is present (Fig. 3-4A, lanes 4 and 5), the mature sized

proteins are replaced by the intermediate forms. This observation implies that the

formation of ml7-protA is dependent on access to the lumen and processing via the ApH-

dependent pathway. At 2.0 tlM pOE23, import of pl7-protA is competitively inhibited,

resulting in inhibition of intermediate processing as well (lane 6). Even at 2.0 [tM

pOE23, all of the imported cpSec pathway precursor, pPC, is processed to mature size.

Following pl7-LF-protA import competition with 0.6 tlM pOE23, fractionation of

chloroplasts was performed to determine the location of products (Fig. 3-4B). All of the

i 7-LF-protA was located in the soluble fraction as would be expected for a stromal

intermediate (lane 5).

Following Processing by the Lumenal Protease, the Mature Substrate Spans the
Thylakoid Membrane with Its Amino Terminus on the Trans Side of the Membrane
and Its Carboxyl Terminus on the Cis Side of the Membrane.

As demonstrated in Fig. 3-2, membrane-associated ml7-protA was sensitive to

thermolysin degradation. Analysis of protease treated membranes on a 12.5% SDS gel

revealed protease protected fragments. The size of the fragment generated depended on

the length and sequence of the linker region. Protease treatment of membrane-associated

ml7-LF-protA yielded two distinct fragments, Fragment 1 and Fragment 2 (Fig. 3-5A,

lane 9). Fragment 1 is approximately 19-kDa; Fragment 2 is approximately 29-kDa (Fig.










3-5B, lanes 3 and 7). Thus, both lumenal fragments are larger than 17-kDa mOE17. As

noted in Fig. 3-2 and Fig. 3-3, much of ml7A-protA is in the stromal fraction. Since the




lonophores

e- c 0



,,. p17A-protA
i 17A-protA

ml7A-protA -
ml7A-protA- .

pl7-L-protA
ml7-L-protA- i 17-L-protA

S-p17-LF-protA
i 17-LF-protA
ml7-LF-protA -

pOE23
. .... .- iO E23
.... .... .. . ...:: ...... ..... ... .. .... i

mO E23 .::,:, :..; ...::..

Lane 1 2 3 4 5 6 7



Figure 3-3. Dissipation of the trans-thylakoid pH gradient during import results in
the accumulation of stromal intermediate forms of 17-protA. pl7A-protA was
generated from rabbit reticulocyte TnT using 35S labeled methionine. pOE23 was
generated from rabbit reticulocyte TnT using 3H labeled leucine. Rabbit reticulocyte
generated translation products were desalted. pl7-L-protA and pl7-LF-protA were
generated from wheat germ TnT using 3H labeled leucine. Translation products (TP, lane
7) were incubated with intact chloroplasts, and ATP in the light at 250C for 10 min in the
absence (lanes 1-3) or presence (lanes 4-6) of nigericin and valinomycin. Chloroplasts
recovered from import assays were subfractionated to membranes and stroma and
visualized by SDS-PAGE (7.5% acrylamide for 17-protA and 12.5% acrylamide for
OE23) and fluorography. In the presence of stromal proteins ml7A-protA migrates as
*ml7A-protA.












Labeled Unlabeled Competitior
A 9-
p17-protA pOE23
[pOE23] [sub-saturating]
[pOE23] iM

TP 0 U .2 .4 .8 2.0
pl7A-protA -il7A-protA il-proaturating]

ml7A-protA -

p17-L-protA a
m17-L-protA- i17-L-protA
pl7-LF-protA w -il7-LF-protA
ml17-LF-protA- -_ -- [pOE23] pM

pOE17- a TP 0 U 2.0
n -iiOE17 pPC- -
mOE17- m m ma mPC- 9flC
Lane 1 2 3 4 5 6 7 Lane 1 2 3 4


B Plus 0.6 pM pOE23
B ------------------------------



Sp17-LF-protA
o- i 17-LF-protA
ml7-LF-protA- f

Lane 1 2 3 4 5 6 7



Figure 3-4. Import of p17-protA was competitively inhibited by ApH-dependent
pathway precursor pOE23. Import substrates pl7A-protA, pl7-L-protA, pOE17, and
pPC were generated in rabbit reticulocyte TnT; p17-LF-protA was generated in wheat
germ TnT. Chloroplasts were preincubated with indicated concentrations of unlabeled
competitor pOE23 and ATP for 7 min in light at 250 C prior to import of labeled
precursors. Incubation was continued 10 min following the addition of labeled import
substrates. (A) Whole chloroplasts were recovered and analyzed by SDS-PAGE (7.5%
acrylamide for 17-protA, 12.5% acrylamide for OE17, and 15% acrylamide for PC) and
fluorography. The lane designated U represents a control assay containing no pOE23
competitor, but containing urea (U) at 167 mM (equal to the concentration of urea in the
2 pM competition assay). (B) Chloroplasts recovered from the import assay were
subfractionated by centrifugation to yield stromal and membrane samples. Samples were
analyzed by SDS-PAGE (7.5% acrylamide) and fluorography.









linker region is only 32 amino acids and the OE17 moiety of pl7A-protA is truncated at

its carboxyl end, OE17 may not fold properly in the lumen and thus be free to back out of

the translocon following processing. Accordingly, protease treated membranes from

pl7A-protA assays yielded little protected peptide (Fig. 3-5A, lane 7). The fragments

retained were approximately the same size as Fragment 1 generated by both ml7-L-protA

(Fig. 3-5A, lane 8) and ml7-LF-protA (Fig. 3-5A, lane 9). The generation of fragments

was dependent on thermolysin treatment, and fragments were not intrinsically protease

resistant. Incubation of membranes without thermolysin, but under otherwise identical

conditions (mock) did not lead to generation of lumenal fragments (Fig. 3-5B, lanes 2 and

6). When recovered membranes were solubilized with Triton-X-100 prior to thermolysin

treatment, both Fragment 1 and Fragment 2 were digested (Fig. 3-5B, lanes 4 and 8).

Therefore, I conclude that the protease resistant fragments are on the trans side of the

thylakoid membrane. Washing with NaOH, but not Na2CO3, disrupts the thylakoid

membrane releasing lumenal contents as seen for mOE23 in lanes 10 and 9 of Fig. 3-5C.

Both protease generated fragments were released by NaOH wash (Fig. 3-5C, lane 5),

indicating that they were soluble inside the lumen.

Fragment 1 and Fragment 2 were derived from OE17 as determined by

immunoprecipitation with appropriate antibodies. As would be expected, pl7-protA

could be immunoprecipitated by irrelevant antibody due to its protein A moiety (Fig. 3-

5D, lanes 6 and 13). Neither Fragment 1 nor Fragment 2 could be immunoprecipitated

by irrelevant antibody (Fig. 3-5D, lanes 4 and 11), but were specifically precipitated by

antibody to OE17 (Fig. 3-5D, lanes 5 and 12). From Fig. 3-5, I conclude that the amino












Membranes



o9 ,< 4 0 <:, 9
z ^\ ^ Y'"
N N N x x x


Membranes


ml 7-L-protA ml7-LF-protA


Frag.--. .


Frag.,-,-


Protease + + +
Lane 1 2 3 4 5 6 7 8 9


Protease mock + + mock + +
Triton X-100 + +
Lane 1 2 3 4 5 6 7 8


Membranes Membranes


TP Membranes TP Membranes
p17-L-protA- .. --
m17-L-protA- -


Fragn-m..I

Fragn.rr i.


Protease - + + +
Na2CO, - + -
NaOH - - +
Lane 1 2 3 4 5


- + + +
- - + -
++ +

- - +
6 7 8 9 10


Fragment 1


Protease
a-Irrelevant
a-OE17
Lane


1 2 3 4 5 6 7 8 9 10 11 12 13 14


Figure 3-5. ml7-protA spans the thylakoid membrane. Analysis of thermolysin
(Protease) treated membranes by SDS-PAGE (12.5% acrylamide) and fluorography
revealed the location and identity of protease protected fragments. (A) pl7-protA was
generated from wheat germ TnT (lanes 1-3). Translation products were imported into
intact chloroplasts. Chloroplasts were recovered and lysed. Recovered membranes were
protease treated with thermolysin as indicated below the panel. (B) pl7-protA was
generated from rabbit reticulocyte TnT. Translation products were imported into intact
chloroplasts. Chloroplasts were recovered and lysed. Recovered membranes were
divided into three aliquots. Membranes were solubilized with 1% Triton X-100 where
indicated below the panels. Samples were protease treated with thermolysin or mock
treated as indicated below the panels. (C) pl7-LF-protA and pOE23 were generated from
wheat germ TnT (TP, lanes 1 and 6). Translation products were imported into intact
chloroplasts. Chloroplasts were recovered and lysed. Recovered membranes were
protease treated with thermolysin as indicated below the panel. Protease treated samples
were subsequently washed with Na2CO3 or NaOH as indicated. (D) pl7-protA was
generated from wheat germ TnT (TP, lanes 1 and 8). Transport assays were conducted
on washed thylakoid membranes and divided into six aliquots. Recovered membranes
were treated with thermolysin as indicated below the panel. Samples in lanes 4, 6, 11,
and 13 were subjected to immunoprecipitation with irrelevant antibody. Samples in lanes
5, 7, 12, and 14 were subjected to immunoprecipitation with anti-OE17.


p17-LF-,:.: ,
m17-LF-i:, :i-


I- jj- i -I

I- jj- i -I


+ + +
+ -+


+ + +
+ -









terminus of pl7-protA entered the lumen where it was processed and retained, while the

carboxyl end remained on the cis side of the membrane.

The Membrane-spanning Intermediate Is Not Integrated into the Lipid Bilayer.

To be of maximum value for investigating ApH-dependent pathway translocation,

the intermediate must be in contact with translocon components. If the ApH-dependent

pathway uses a channel, it can be imagined that ml7-protA remains in the channel

spanning the membrane. Alternatively, the processed protein may have slipped out of the

channel and been integrated into the lipid bilayer. The latter possibility was a particular

concern given the presence of two mildly hydrophobic stretches of amino acids in the

linker detailed in Fig. 3-1B. The size of Fragment 1, approximately 19-kDa, implies that

membrane-associated ml7-L-protA rests primarily in the region derived from

dihydrofolate reductase (DHFR). The appearance of Fragment 2 from ml7-LF-protA

demonstrates that the change in the linker region allowed additional amino acids to reach

the lumen. The linker region of pl7-LF-protA includes a highly charged region derived

from Kodak FLAG that effectively lowers the hydrophobicity of the DHFR segment.

The size of Fragment 2 could reflect either translocation up to the protein A moiety or

arrest in the relatively hydrophobic region derived from the small subunit of RUBISCO.

Arrest by hydrophobic segments of the linker region may be dictated by association with

membrane proteins or with the bilayer. A cartoon illustrating the proposed derivation of

Fragments 1 and Fragment 2 from ml7-LF-protA is displayed in Fig. 3-6C.

To test for bilayer integration, I imported p 7-protA, recovered the membranes,

and washed them with IB, NaC1, Na2CO3, NaOH, or urea (Fig.3-6A). The membrane-







54


associated ml7-protA was fully resistant to all treatments except NaOH (lane 5). Only

about 40% of membrane-associated pl7-protA was resistant to NaOH wash (Fig. 3-6D).

A oB Membranes

TP0 0"Z C V V

pl7-L-protA mm <
m17-L-protA- O 1 R l

p17-LF-protA- I Fragment 2
m17-LF-protA- Fragment 1
Lane: 1 2 3 4 5 6 7 8
Protease - + + + +
C Lane: 1 2 3 4 5 6 7 8 9 10 11 12
protein A
alanine 290


protein A
Protease 2. Protease


valineanie
alanne 290 alanne 290





0E17 valine 229

Fragment 2



Figure 3-6. The mildly hydrophobic stretches of amino acids in the linker region of
pl7-protA do not cause the translocation intermediate to move into the lipid bilayer.
(A) Intact chloroplasts were incubated with pl7-protA from rabbit reticulocyte TnT (TP,
lane 1) and ATP in light at 250C for 15 min. Chloroplasts were recovered and lysed.
Recovered membranes were washed as indicated above each lane. Samples were
analyzed by SDS-PAGE (7.5% acrylamide) and fluorography. (B) pl7-LF-protA and
three similar precursors having differences in one or two amino acids in the linker region
were generated in wheat germ TnT (lanes 1-4). Intact chloroplasts were incubated with
precursors and ATP in light at 250C for 15 min. Chloroplasts were recovered and lysed.
Recovered membranes were protease treated with thermolysin as indicated below the
panel. Samples were analyzed by SDS-PAGE (12.5% acrylamide) and fluorography.
(C) Models of ml7-LF-protA membrane-spanning intermediate are displayed. For
clarity, the length of the linker region is greatly exaggerated relative to the protein A
moiety and OE17. Colors in the linker region correspond to colors in Fig. 3-1 B and C.
















IB Washed Membranes


NaOH
Washed
Membranes


NaOH
IB Washed Washed
Membranes Membranes


Dilution: 1:1 1:2 1:4 1:8 1:1 1:2 1:1 1:2 1:1


ml7-LF-protA- W

mLF(V/R)(A/R) -0


mLF1- W


w

m -- NA


U a mLHCP


Standard Curve: Least Squares
for mLF(V/R)(A/R)


Membrane-
associated
Protein

ml7-LF-protA

mLF(V/R)(A/R)

mLF1

mLHCP


Estimated Percentage
Resistant to
NaOH Wash


.m4

.11"..4


II I J3OH .

/
1:4, IB


S 'l 2, NaOH
, 1:8, IB


ni *E pi afr a
Percentage of Standard Sample



Figure 3-6. The mildly hydrophobic stretches of amino acids in the linker region of
pl7-protA do not cause the translocation intermediate to move into the lipid bilayer
(continued). (D) Intact chloroplasts were incubated with pLHCP generated from mRNA
and wheat germ or pl7-protA from wheat germ TnT and ATP for 30 min in light at 250C.
Chloroplasts were repurified, lysed, and recovered membranes were washed as indicated
above the panels. Membranes were again recovered and samples were adjusted to equal
chlorophyll content, diluted as indicated above the panels, and analyzed by SDS-PAGE
(12.5% acrylamide). The density of scanned bands from the X-ray film was determined
using Alpha Imager software. Relative protein quantities were estimated by averaging
values derived from 1:1 and 1:2 dilution points of NaOH washed membranes.




From these data, I conclude that the membrane-spanning intermediate is not integrated in


the bilayer.


To further test the influence of the amino acid sequence in the linker region, I


protease treated membranes following import of three mutated versions of pl7-LF-protA









diagramedd in Fig. 3-1B). Within a stretch of fifteen amino acids derived from DHFR

identified as mildly hydrophobic, the eighth amino acid (valine 229) was mutated to

arginine. The change of valine 229 to arginine in pLF(V/R) resulted in an increased

abundance of Fragment 2 relative to Fragment 1 (Fig. 3-6B, lane 10). Within a stretch of

ten amino acids derived from the small subunit of RUBISCO identified as mildly

hydrophobic, the fourth amino acid (alanine 290) was mutated to arginine. The change of

alanine 290 to arginine in pLF(A/R) had no apparent influence on the relative abundance

of Fragment 2 versus Fragment 1 (Fig. 3-6B, lane 11). When both valine 229 and alanine

290 were changed to arginine in pLF(V/R)(A/R) Fragment 2 dominated (lane 12). The

variations in size and abundance of the fragments generated means that the nature of the

amino acids in the linker region influences the point at which the majority of intermediate

rests relative to the cis side of the membrane. However, mutating hydrophobic amino

acids to positively charged arginine did not prevent arrest. Because changing alanine 290

to arginine did not increase the apparent size of Fragment 2, it is likely that Fragment 2

results from translocation up to the beginning of the Protein A moiety. In other

experiments, I imported pLF(V/R)(A/R), recovered the membranes, and washed them

with NaOH. As with ml7-LF-protA, about 60% of mLF(V/R)(A/R) was released (Fig.

3-6D). Therefore, it does not appear that the hydrophobicity of the amino acids in the

linker region determines translocation arrest.

The Membrane-spanning Intermediate is Arrested Due to the Protein A Moiety.

As seen in Fig. 3-6, mLF(V/R)(A/R) is translocated up to the beginning of the

protein A segment. The reason for arrest is unclear, given that folded protein transport

has been documented on the ApH-dependent pathway and the related bacterial Tat









pathway. I therefore investigated the size of protein A segment necessary to arrest

translocation by importing three truncated versions of pl7-LF-protA: LF1, LF2, and LF3

having one, two, or three IgG binding domains, respectively (Fig. 3-1C). Following

import, I recovered the membranes and protease treated with them thermolysin. Each

substrate generated Fragment 1 and Fragment 2 as seen for full-length pl7-LF-protA

(Fig. 3-7A, lanes 9-12).

I demonstrated that arrest is dependent on at least one IgG binding domain by

removing the entire protein A moiety enzymatically and by genetic truncation. The

genetically truncated precursor, pLFO is processed and exported to the thylakoid lumen

where it is resistant to externally added protease. The mature product is labile to

endogenous proteases so that it could only be observed by arresting lumenal protease

activity with HgC12 prior to thermolysin treatment. A time-course revealed accumulation

and endogenous degradation of mLFO (Fig. 3-7B).

The linker region of pl7-L-protA includes an engineered thrombin cleavage site

(See Fig. 3-1B for amino acids and Fig. 3-6C, lavender segment, for illustration). Partial

digestion of the precursor with thrombin prior to import assays yielded three products:

undigested pl7-L-protA, the amino segment up to the thrombin site (p17-T), and the

carboxyl segment beyond the thrombin site (T-protA) (Fig. 3-7C lane 4). Because T-

protA has no stromal targeting domain, it did not enter the chloroplasts during the

subsequent import assay. The amino end of the precursor, pl7-LT, was imported,

processed to ml7-LT, and fully transported to the lumen (Fig. 3-7C, lane 5). Treatment

of recovered membranes with thermolysin demonstrated that ml7-LT is on the trans side

of the membrane (Fig. 3-7C, lane 6).








58





Membranes



, <



Fragment 2
Fragment 1


Protease - - + + +
Lane: 1 2 3 4 5 6 7 8 9 10 11 12


r I I I i









Thrombin
Thremolysin
Lane: 1 2


TP 0.5 1.0 1.5 2.0 2.5 3.0

.


- p17-L-protA


-T-protA
-p17-LT

m1l7-LT


...I |


3 4 5 6


4.0 5.0 7.0 10.0 Minutes
pLFO
-mLFO


rmr I


Figure 3-7. Translocation arrest is dependent on the Protein A segment of pl7-
protA. All samples were analyzed by SDS-PAGE (12.5% acrylamide) and fluorography.
(A) Precursor proteins generated from wheat germ TnT (lanes 1-4) were incubated with
intact chloroplasts and ATP for 15 min in light at 250C. Chloroplasts were recovered and
lysed. Recovered membranes were protease treated with thermolysin where indicated
below the panel. (B) Washed thylakoid membranes were incubated with precursor
protein generated from wheat germ TnT in light at 250C. Samples were taken at the times
indicated above the panel and added to tubes containing HgC12 (final concentration 1.6
mM). Membranes were recovered, HgC12 was removed by washing with IB containing
EDTA, and samples were protease treated with thermolysin. (C) pl7-L-protA (TP, lane
1) was generated in wheat germ TnT; an aliquot was partially digested with thrombin
(TP, lane 4). Intact chloroplasts were incubated with ATP and translation product
(native, lane 2 or thrombin treated, lane 5) and ATP for 15 min in light at 25oC.
Chloroplasts were recovered and lysed. Recovered membranes were protease treated
with thermolysin where indicated below the panel.


Membranes


Membranes









The Protein A Moiety Alone Is Insufficient to Arrest Transport.

The data exhibited in Fig. 3-6 demonstrated that arrest was not due to bilayer

integration resulting from mildly hydrophobic stretches in the linker. The data exhibited

in Fig. 3-7 demonstrated that arrest was dependent of the presence of a least one IgG

binding domain carboxy to the linker. Thus, the question arises: Is arrest dependent on

some inherent property of the IgG binding domains, or is it the result of the position of

those domains? In pursuit of an answer to that question, the full protein A derived

moiety was linked carboxy to the complete transit peptide from pOE17 without an

intervening mOE17 protein or linker, yielding pl7tp-protA.



Fractionation of Import





TP
pl7tp-protA- ,.
il7tp-protA -
i1 7tp-protA ml7tp-protA


Protease + mock + mock +

Lane: 1 2 3 4 5 6 7 8


Figure 3-8. The ApH-dependent pathway is able to transport protein A. pl7tp-
protA generated from wheat germ TnT was incubated with intact chloroplasts (Import) or
lysate (transport) and ATP for 15 min in light at 250C. Chloroplasts were recovered from
the import assay and subfractionated to yield stromal and membrane samples. Recovered
membranes were protease treated with thermolysin or mock treated as indicated beneath
the panel. Samples were analyzed by SDS-PAGE (12.5% acrylamide).









The construct, pl7tp-protA, was translated in vitro and incubated with

chloroplasts under import permissive conditions. As exhibited in Fig. 3-8, pl7tp-protA

was imported and processed to an intermediate and to a mature form. A small amount of

intermediate remained in the stroma (lane 4); however, most of the translation product

was processed to mature size and localized to the thylakoid lumen where it was resistant

to externally applied protease (lane 6). A transport assay using lysed chloroplasts

resulted in processing to mature size (lane 7) with most of the processed substrate

resistant to externally applied protease (lane 8). Thus, it is not an inherent property of the

protein A derived moiety of pl7-protA that arrests transport on the ApH-dependent

pathway.


Discussion

The ApH-dependent pathway and the related bacterial Tat pathway are unique

among export systems in being able to translocate folded proteins. The mechanism of

protein transport and energy utilization is unknown for either system. Numerous

laboratories have attempted to create a membrane-spanning translocation intermediate in

the thylakoid system. Attempts have resulted in two published accounts of folded protein

translocation. Clark and Theg (1997) demonstrated that the tightly folded 6.5-kDa

bovine pancreatic trypsin inhibitor (BPTI) linked behind the ApH-dependent pathway

precursor to OE17 (pOE17) could be fully translocated to the thylakoid lumen. Transport

proceeds even when the BPTI moiety, estimated to be 2.3 nm in diameter, is internally

crosslinked and therefore incapable of unfolding during translocation. Evidence of the

system's ability to export a somewhat larger folded domain came from a study in which









dihydrofolate reductase (DHFR) was linked to pOE23 and shown to enter the lumen even

when bound to methotrexate (Hynds et al., 1998).

Two laboratories have achieved translocation arrest. A biotinylated protein,

i23K-BCCP, completed to avidin was processed, but the bulk of the protein apparently

remained on the cis side of the membrane, since no lumenal fragment was recovered

(Asai et al., 1999). When i23K-BCCP is completed with avidin prior to incubation with

isolated thylakoids, the mature-sized product is formed indicating processing by the

lumenal protease; however, the m23K-BCCP remains sensitive to thermolysin. These

results imply that export has been arrested with the amino end of the substrate protein in

the lumen while most of the protein remains on the cis side of the membrane. Additional

data from the laboratory that achieved the arrest has not been forthcoming. A fusion

protein having the lumenal targeting domain of OE17 and the mature domain of OE23

was partially translocated after being abnormally processed, but the amino end, not the

carboxyl end, was apparently retained in the membrane (Berghofer and Klosgen, 1999).

The nature of its association with the membrane was unknown.

I have generated a membrane-spanning translocation intermediate, confirmed its

transport via the ApH-dependent pathway by energetic and competition studies, and

investigated the nature of its membrane association. I have confirmed that the protein A

derived domain of the ml7-protA is responsible for translocation arrest; however, the

reason for that arrest is unclear, as the ApH-dependent pathway is able to transport the

protein A moiety. A possible cause for arrest ml7-protA versus complete transport of

ml7tp-protA is the presence of the linker region (seventy amino acids) between the

mature OE17 precursor and the protein A derived domain.









Evidence has mounted for initial interaction of precursor with an Hcfl06/cpTatC

complex and subsequent interaction with Tha4 (Ma and Cline, 2000; Cline and Mori,

unpublished; Mori et al., submitted). The isolation of a membrane-spanning translocation

intermediate supports the idea of channel utilization by the ApH-dependent pathway. It is

possible that channel formation is a dynamic process with pore size being determined

early. If so, then translocation of folded mature OE17 followed by an unfolded linker

might allow adjustment of the channel, since the ion impermeable state of the membrane

must be maintained. In the absence of the transit peptide, it may not be possible for the

channel size to be re-adjusted to accommodate the folded protein A moiety following the

linker. Future research may take advantage of this tool to elucidate the mechanism of

translocation and the participation of pathway components.














CHAPTER 4
MEMBRANE INTEGRATION OF IN VITRO-TRANSLATED ApH-DEPENDENT
PATHWAY COMPONENTS




Introduction

The Tat operon of E. coli encodes three proteins known to function in Tat

dependent protein translocation: TatA, TatB and TatC (Weiner et al., 1998; Sargent et al.,

1998). A fourth gene encodes a protein, TatE, similar in sequence and under some

conditions functionally interchangeable with TatA (Sargent et al., 1998). An orthologue

to TatB was first discovered in maize and designated Hcfl06 (Voelker and Barkan,

1995). An orthologue of TatA/E was discovered in pea and termed Tha4 (Mori et al.,

1999). Subsequently, two orthologues of TatA/E were discovered in maize and termed

Tha4 and Tha9 (Walker et al., 1999). The cDNAs for maize Hcfl06 (Settles et al., 1997)

and Tha4 (Walker et al., 1999) have been cloned, as have been cDNAs for Hcfl06 (Mori

et al., submitted), Tha4 (Mori et al., 1999), and cpTatC (Mori et al., submitted) from pea.

The precursor to Tha4 (pTha4), transcribed and translated in vitro from pea

cDNA, was imported into chloroplasts where it was processed to mature size (mTha4)

and localized to the thylakoid membrane (Mori et al., 1999). The thylakoid-associated

mTha4 was resistant to Na2CO3 wash, but protease sensitive. In submitted work, Mori et

al. performed similar experiments on precursors for cpTatC (pcpTatC from pea) and

Hcfl06 (pHcfl06 from both maize and pea). Import of the 36-kDa pcpTatC resulted in a

34-kDa mature membrane-associated form (mcpTatC). mcpTatC was resistant to









washing with Na2CO3, but partially degraded by thermolysin yielding fragments of 26-

kDa and 23-kDa.

Blue native polyacrylamide gel electrophoresis (BN PAGE) and immunoblotting

with antibodies to cpTatC and to Hcfl06 demonstrated association of those two

endogenous components in a 700-kDa complex (Mori et al., submitted). Smaller species

of Hcfl06 were observed from a diffuse band just below 700 kDa to an approximately

80-kDa band depending on detergent/protein ratio. Tha4 was detected in bands from

about 400 kDa to about 70 kDa, also depending on detergent/protein ratio. However, no

conditions were found under which Tha4 appeared at the same position as Hcfl06 or

cpTatC.

The putative transmembrane domains of Tha4 and Hcfl06 include a conserved

glutamate. Because there is an energetic cost to maintaining a negatively charged amino

acid in a transmembrane domain, the conservation of glutamates in these two components

is likely to be of functional significance. The nature of the association between

endogenous cpTatC and Hcfl06 is unknown. One possible interaction may take place

between the conserved glutamate in the transmembrane domain of Hcfl06 and a

conserved arginine in a transmembrane domain of cpTatC.

Biochemical analysis of protein translocation has been most effectively pursued

using isolated chloroplasts from peas with in vitro-translated translocation substrates.

The cloning of pathway components Hcfl06, Tha4, and cpTatC has made their gene

products available for in vitro mutation, antibody production and for in vitro translation.

I have investigated the integration of in vitro-translated components as translocation

substrates in thylakoid membranes of pea and maize. Integration of radiolabeled









components allows the visualization of complex formation by fluorography. The effects

of changes in amino acids within components on complex formation can thus be

monitored. The position of complexes can be visualized, allowing a shift in the complex

size due to translocation substrate binding to be investigated. One long-term goal is the

reconstitution of mutant maize thylakoid membranes to wild type using in vitro-translated

components. Reconstitution would be a step in the definition of component roles.


Materials and Methods

Preparation of Precursor Proteins

Cloning and analysis of DNA products were by standard molecular biology

procedures (Sambrook et al., 1989). Amplifications were performed with Pfu polymerase

(Stratagene, La Jolla, CA). Cloned constructs were verified by DNA sequencing.

Sequencing was done with ABI Prism Dye Terminator cycle sequencing protocols

developed by Applied Biosystems (Perkin-Elmer Corp., Foster City, CA) and an Applied

Biosystems model 373 Stretch DNA Sequencer (Perkin-Elmer Corp.). Sequencing of all

clones on both strands was performed by the University of Florida Interdisciplinary

Center for Biotechnology Research (ICBR) DNA Sequencing Core Facility.

The mature form of zm Hcfl06 (zm mHcfl06) was cloned by PCR amplification

from zm pHcfl06 (Settles et al., 1997) based on the transit peptide cleavage site predicted

by ChloroP (Emanuelsson et al., 1999)1. The 5' primer (including an engineered EcoRI

site) was used to mutate the nucleotides encoding cysteine 67 to encode methionine; the

3' primer bound in the pGEM 4Z vector. The resulting product was ligated into pGEM

4Z at the EcoRI and BamHI sites in the SP6 direction. The mature form of ps Hcfl06


1zm mHcfl06 was cloned by Hiroki Mori in 1998.









(ps mHcfl06) was cloned by PCR amplification from ps pHcfl06 (Mori et al. submitted)

based on the transit peptide cleavage site predicted by ChloroP (Emanuelsson et al.,

1999)2. The 5' primer (including an engineered EcoRI site) was used to mutate the

nucleotides encoding tyrosine 86 to encode methionine; the 3' primer bound in the

pGEM 4Z vector. The resulting product was ligated into pGEM 4Z at the EcoRI and SstI

sites in the SP6 direction. An altered from ofps mHcfl06, ps mHcfl06 E/Q, was derived

by PCR amplification using a 5' primer (including an engineered KpnI site) that mutated

nucleotides encoding glutamate 11 to glutamine and a 3' primer that bound within pGEM

4Z3. The resulting product was ligated into pGEM 4Z at the KpnI site in the SP6

direction.

The mature form of ps Tha4 was cloned by PCR amplification from ps pTha4

(Mori et al., 1999) based on the transit peptide cleavage site predicted by ChloroP

(Emanuelsson et al., 1999)4. The 5' primer (including an engineered KpnI site) was used

to mutate the nucleotides encoding aspargine 56 to encode methionine; the 3' primer

bound in the pGEM 4Z vector. The resulting product was ligated into pGEM 4Z at the

KpnI site in the SP6 direction. An altered form of ps mTha4, ps mTha4 E/Q, was derived

by PCR amplification using a 5' primer (including an engineered KpnI site) that mutated

nucleotides encoding glutamate 10 to glutamine and a 3' primer that bound within pGEM

4Z5. The resulting product was ligated into pGEM 4Z at the KpnI site in the SP6

direction.




2ps mHcfl06 was cloned by Hiroki Mori in 1998.
3 ps mHcfl06 E/Q was cloned by Mike McCaffery in 1999.
4 ps mTha4 was cloned by Hiroki Mori in 2000.
5ps mTha4 E/Q was cloned by Hiroki Mori in 2000.









The mature form of ps TatC was cloned by PCR amplification from ps pTatC

(Mori et al., submitted) based on the transit peptide cleavage site predicted by ChloroP

(Emanuelsson et al., 1999)6. The 5' primer (including an engineered EcoRI site) was

used to mutate the nucleotides encoding leucine 39 to encode methionine; the 3' primer

bound in the pGEM 4Z vector. The resulting product was ligated into pGEM 4Z at the

EcoRI and BamHI sites in the SP6 direction.

Preparation of Radiolabeled Precursors

In vitro transcription with SP6 RNA polymerase (Promega) and translation with

wheat germ lysate (Promega) or coupled transcription/translation with wheat germ TnT

(Promega) in the presence of 3H leucine (NEN Life Science Products) was performed

following the manufacture's guidelines. Translation products were diluted with one

volume 60 mM leucine in 2X import buffer (IX = 50 mM HEPES, KOH, pH 8.0, 0.33 M

sorbitol) prior to use unless otherwise indicated in the figure legend.

Preparation of Chloroplasts, Thylakoids, and Lysate

Intact chloroplasts were isolated from 9- to 10-day old pea seedlings (Pisum

sativum cv. Laxton's Progress 9) as described (Cline et al., 1993) and were resuspended

in import buffer (IB). Maize plants were grown at 260 C in a 16 h light/8h dark cycle for

7-10 days. Experiments on exclusively wild-type maize were conducted on cv. Trucker's

Favorite. Mutant hcf]06 mum3 maize seedlings were selected by their pale green

phenotype and confirmed by high chlorophyll fluorescence with a hand-held UV lamp.

Experiments using hcf]06 mum3 chloroplasts were controlled by wild-type chloroplasts

isolated from the same cv. Maize chloroplasts were isolated by essentially the same

procedure as pea chloroplasts except that intact chloroplasts were purified on a step


6 ps mTatC was cloned by Hiroki Mori in 2000.









density gradient consisting of 10 ml of 75% Percoll and 25 ml of 35% Percoll in GR

buffer lacking both magnesium and manganese ions. The gradients were centrifuged at

3200 X g for 15 min, and intact chloroplasts were collected at the 35%/75% interface.

Lysate and washed thylakoids were prepared from isolated chloroplasts (Cline et al.,

1993). Chlorophyll concentrations were determined according to Amon (1949).

Chloroplast Import and Thylakoid Protein Integration Assays

Import of radiolabeled precursors into isolated chloroplasts or integration into

washed thylakoids or chloroplast lysate was conducted in microcentrifuge tubes in a 25C

water bath illuminated with 70 [tE m-2s-1 white light (Cline et al., 1993) for 10 min or the

time indicated in the figure legend. Assays were generally terminated by transfer to 0C.

Where indicated, recovered chloroplasts or thylakoids were protease post-treated with

thermolysin. Chloroplasts were repurified on Percoll cushions; thylakoids were

recovered by centrifugation.

Quantitative Immunoblots

Immunoblots were developed by ECL procedure (Pierce). For quantitation of

translation products, translation products were run on SDS-PAGE in parallel with

dilution series ofHcfl06 stromal domain or Tha4 stromal domain standards. The content

of standard solutions was determined by amino acid analysis conducted by the University

of Florida Interdisciplinary Center for Biotechnology Research protein core facility.

Proteins were electroblotted to nitrocellulose membranes and then immunodecorated with

the appropriate antibodies. The density of scanned bands on X-ray film was determined

using Alpha Imager software and protein quantities were estimated by comparison to

standards in the linear region of the film.









Blue Native Polyacrylamide Gel Electrophoresis

Blue native polyacrylamide gel electrophoresis (BN PAGE) was carried out as

described by Schagger and von Jagow (1991) with the following modifications. Washed

thylakoids were suspended in resuspension buffer (20% glycerol, 25mM BisTris-HC1, pH

7.0) at 2.0 mg chlorophyll/ml. An equal volume of resuspension buffer containing twice

the final digitonin concentration was added while gently vortexing. After incubation at

4C for 30 min with end over end mixing, insoluble material was removed by

centrifugation at 200,000 x gave for 20 min. The resulting supernatant was combined with

1/10 volume of 5% Serva Blue G, 100 mM BisTris-HC1, pH 7.0, 0.5 M 6-amino-n-

caproic acid, 30% sucrose and subjected to electrophoresis through 8.4 cm x 5 cm x 0.75

mm mini gels in a Hoefer Mighty Small vertical electrophoresis unit connected to a

cooling circulator. The separating gel consisted of a linear 5%-13.5% acrylamide

gradient and 5-15% glycerol gradient. The stacking gel was 4% acrylamide. Blue native

gels were run at 100-200 V for 3-4 hrs at 2-40C. The cathode buffer was exchanged with

cathode buffer lacking dye after the top of 1/2-1/3 of the gel was covered with dye and

electrophoresis was continued until free dye cleared the bottom of the gel. Gels to be

analyzed for fluorography were then treated with DMSO and PPO as described.

Molecular markers used for blue native gels are ferritin (880 kDa and 440 kDa) and

bovine serum albumin (132 kDa and 66 kDa).


Results

In Vitro-translated ApH-dependent Pathway Components Associated with
Thylakoid Membranes.

As has been previously reported, in vitro-translated components were imported

into intact chloroplasts and processed to mature size. Fig. 4-1A exhibits the relative sizes









of precursors and integrated components. The figure is taken from a single gel divided

into panels for clarity, but with the relative location of bands maintained to reflect

apparent sizes. Also displayed in Fig. 4-1A and Fig. 4-1B are assays performed with

mHcfl06, mcpTatC, and mTha4 generated from transcription plasmids lacking the

coding regions for the transit peptides. Association of translated mcpTatC with

membranes from lysate was inefficient as compared to association following import and

in organello processing (Fig. 4-1A, lane 6 versus lane 7). Translated mHcfl06 exhibited

very efficient membrane association in lysate as compared to imported and processed

mHcfl06 (Fig. 4-1A, lane 4 versus lane 2).

It has been previously determined that following import assays of precursor

components, processed-cpTatC and Hcfl06 are resistant to Na2CO3 wash, indicating tight

membrane association (Mori et al., submitted). I tested for membrane integration using

the more stringent criterion of resistance to NaOH wash. Integration of membrane

protein cpSecY was used as a control. Some mature sized cpTatC was recovered in the

stromal fraction (Fig. 4-1B, top left panel, lane 3). The membrane-associated mcpTatC

is resistant to NaOH wash (Fig.4-1B, top left panel, lane 6) and (as previously reported)

yields two protease-protected fragments (Fig. 4-1B, top panel, lane 7). In vitro-translated

mHcfl06 integrated into isolated thylakoids as demonstrated by resistance to NaOH wash

(Fig. 4-1B, upper right panel, lane 3). Mutation of a glutamate to glutamine in the

putative transmembrane domain of Hcfl06 did not abrogate integration as measured by

resistance to NaOH wash (Fig. 4-1B, upper right panel, lane 6). Mori et al. (1999)

reported that following import, the membrane-associated mTha4 is resistant to Na2CO3

wash. In unpublished work, Summer and Cline observed that endogenous Tha4 is













pHcfl06 mHcfl06 mcpTatC pcpTatC



TP TP P TP TP


Lane: 1 2 3 4 5 6 7 8


0- Membranes

TP --------


mHcfl06 mHcfl06 E/Q

TP Membranes TP Membranes


NaOH


Lane: 1 2 3 4 5 6



pTha4 mTha4 mTha4 E/Q

TP Membranes TP Membranes TP Membranes


mock +

mock - +
2 3 4 5 6 7


NaOH


Lane: 1 2 3 4 5 6 7 8 9


Figure 4-1. In vitro-translated ApH-dependent pathway components associated with
thylakoid membranes. All substrates were generated from wheat germ TnT. (A)
Translated component-substrates were incubated with intact pea chloroplasts (Import) or
lysate (Integration) and ATP for 15 min in the light at 250C. Chloroplasts were lysed;
membranes were recovered and washed with IB. Samples were analyzed by SDS-PAGE
(12.5 % acrylamide) and fluorography. (B) Left panels: Intact chloroplasts were
incubated with pcpTatC or pSecY and ATP for 30 min in the light at 250C. Chloroplasts
were recovered; aliquots were lysed and subfractionated by centrifugation to yield stroma
and membranes. Membrane fractions were washed with NaOH, treated with thermolysin
(protease), or mock treated with IB as indicated. Samples were analyzed by SDS-PAGE
(11 % acrylamide) and fluorography. Right panels: Component-substrates were
incubated with lysate and ATP at 250C for 15 min in the light. Membranes were
recovered. Samples were divided and washed with NaOH or with IB as indicated.
Samples were analyzed by SDS-PAGE (12.5 % acrylamide) and fluorography.


Translation
Product:


pHcfl06 -

mHcfl06 -


pTha4


TP


pcpTatC
mcpTatC


r,.Tl,. .-


9 10 11


pcpTatC -


pSecY -


- mcpTatC

- DP1 cpTatC
- DP2 cpTatC


- mSecY


NaOH

Protease
Lane:


- DP SecY


-









sensitive to NaOH wash as measured by subsequent immunoblotting of membranes with

antibody to Tha4. Incubation of in vitro-translated pTha4 with lysate at 25 o C, 5mM

ATP, and actinic light for 15 min resulted in processing to mature size and membrane

association (Fig. 4-1B, lower right panel, lane 2). The membrane-associated mTha4 was

released by NaOH wash (lane 3). The same assay and wash when performed on in vitro-

translated mTha4 and mTha4 E/Q (glutamate in putative transmembrane domain mutated

to glutamine) also resulted in release of membrane-associated label (lanes 5, 6, 8, and 9).

Therefore, while the possibility of the hydrophobic domain of Tha4 integrating into the

thylakoid membrane cannot be excluded, the association is not as strong as can be

demonstrated for Hcfl06 and cpTatC.

Quantitative immunoblotting was used to estimate the amount of translation

product added to assays (Fig. 4-2A). Comparison of density scans of radiolabeled

translation product with integrated radiolabeled translation product allowed me to make

an estimate of the amount of translated mHcfl06 integrated in membranes (Fig. 4-2B and

4-2C). Approximately 2.0 x 10-13 moles mHcfl06 translation product were integrated per

|tg chlorophyll. A previous analysis of endogenous Hcfl06 (Mori et al., submitted) by

quantitative immunoblotting estimated 1.6 x 10-13 moles per |tg of chlorophyll.

Quantitation of membrane-associated mTha4 was accomplished as described for

mHcfl06. About 8.5 x 10-13 moles of translation product per |tg of chlorophyll were

membrane-associated (Fig. 4-3). A previous estimate of endogenous Tha4 found about

2.3 x 10-13 moles per |tg chlorophyll (Mori et al., submitted).

The integration of in vitro-translated components into the thylakoid membrane in

quantities comparable to endogenously present components means that it may be feasible







73



. Dilution of Standard X 10"

4.0 3.0 2.0 1.5 1.0




2 3 4 5 6


Membranes
after IB Wash

Dilution: 1:4 1:8 1:16 1:32


Point Number: 2 3 4 5


Standard Curve: Least Squares
for Stromal Domain Standard of Hcfl06







Di "o o TP@1:16



D 1o1 Sit J 11 IT
Dilution of Standard X 10"


Membranes
after NaOH Wash Translation Product

1:4 1:8 1:16 1:32 1:20 1:40 1:80 1:160


6 7


8 9 10 11 12 13


Standard Curve: Least Squares
for Translation Product mHcfl06


>,
m


c n mn


0


T'l
1 ii 1 11 111-





Estimated Translation Product
x 10-14 Moles per Lane


Figure 4-2. Quantitation of in vitro-translated mHcfl06 integrated into thylakoid
membranes. Substrate mHcfl06 was generated from mRNA and wheat germ extract.
(A) A quantitative immunoblot of translation product was developed using expressed
stromal domain ofHcfl06 as a protein standard. (B) In vitro-translated mHcfl06 was
incubated with lysate and ATP under actinic light for 10 min at 250C. Membranes were
recovered and washed as indicated. Samples were analyzed by SDS-PAGE (12.5%
acrylamide) and fluorography. (C) The density of scanned bands from the X-ray film
pictured in A an B was determined using Alpha Imager software. The standard curve
was drawn from points in yellow; the values of unknowns in the linear range were read
from points in white.


Point Number:


404 ..







74


Dilution of
A Standard 0
<- X10-4 g-
--------------------
5.0 7.0 9.0 '



Point Number: 1 2 3 4 5
B
Membranes Membranes
after IB Wash after Na2CO3 Wash Translation Product

Dilution: 1:4 1:8 1:16 1:32 1:4 1:8 1:16 1:32 1:10 1:20 1:40 1:80 1:160



Point Number: 1 2 3 4 5 6 7 8 9 10 11 12 13


C
Standard Curve: Least Squares Standard Curve: Least Squares
for Stromal Domain Standard of Tha4 for Translation Product mTha4


a ZMS TP@1:16 a-) I '"




ro 4 Q t of td

5, a e o m d hea n3r E (
> 4 / n aj. f






membranes Substrate mTha4 was generated from mRNA and wheat germ extract. (A)







A quantitative immunoblot of translation product was developed using expressed stromal
domain of Tha4 as a protein standard. (B) In vitro-translated mTha4 was incubated with
lysate and ATP under actinic light for 10 min at 25C. Membranes were recovered and
washed as indicated. Samples were analyzed by SDS-PAGE (12.5% acrylamide) and
fluorography. (C) The density of scanned bands from the X-ray film pictured in A an B
was determined using Alpha Imager software. The standard curve was drawn from
points in yellow; the values of unknowns in the linear range were read from points in
white.
doano h4a rti sadr.()I ~otasatdmh4wsicbtdwt





white.fSadadX104Etmae rasain rdc









to evaluate the effects of changes in the amino acid sequence of those components on

complex formation. The integration of altered components into endogenous complexes

could alter the physiological performance of the complexes. In mutant membranes

lacking a given component, it may be possible to integrate sufficient amounts of in vitro-

translated protein to restore translocation competence.

Association of In Vitro-translated Components with Endogenous Complexes.

It was previously determined that endogenous Hcfl06 and cpTatC are part of a

700-kDa complex (Mori et al., submitted). Import into intact chloroplasts of in vitro-

translated pcpTatC resulted in its integration into an approximately 700 kDa complex that

was then visualized by BN PAGE and fluorography (Fig. 4-4 A, lanel; 4-4B, lanel; and

4-4C, lanel). Integration of mcpTatC was inefficient, but also resulted in association

with the 700-kDa complex (Fig. 4-4A, lane 3).

Import into intact chloroplasts of pHcfl06 (Fig. 4-4B, lane 2 and 4-4C, lane 2) or

mHcfl06 (Fig. 4-4C, lanes 5, 6, and 14) resulted in integration into a 700-kDa complex.

In the presence of 1% digitonin, a separate pool of Hcfl06 was visualized at lower

molecular weights (Fig. 4-4C, lanes 5, 6, and 14). Integrated mHcfl06 E/Q did not

associate with the 700-kDa complex (Fig. 4-4C lanes 8 and 15), but did generate lower

molecular weight species. Similar lower molecular weight species could be generated by

combining mHcfl06 or mHcfl06 E/Q translation product with solubilized membranes

(Fig. 4-4C, lanes 12 and 13). The implications from these data are that only the 700-kDa

complex may be physiologically relevant and that when the glutamate in the

transmembrane domain of mHcfl06 is mutated to glutamine, the component is no longer

able to associate in the 700-kDa complex.









Import of pTha4 into intact chloroplasts (Fig. 4-4 A, lane 2 and 4-4C, lane 3) or

integration of pTha4 or mTha4 into membranes from lysate resulted in species migrating

from about 100 kDa (Fig. 4-4A, lane 4) to about 400 kDa (Fig. 4-4B, lane 3 and 4-4C

lanes 3, 7, and 20) depending on the amount of digitonin present and on the amount of

sample loaded on the gel. Integration of mTha4 E/Q into membranes from lysate yielded


Approximate Size


Import

-O
6^A$


Integration
-C-


700 kDa -


400 kDa -


Import

----- ---- -


880 kDa -

440 kDa -


130 kDa -
100 kDa -


Lane:


1 3
1 2 3


Lane: 1 2 3 4


Figure 4-4. Incorporation of in vitro-translated components into native complexes.
Substrates were generated from wheat germ TnT. Samples were analyzed by BN PAGE
and fluorography. (A) Chloroplasts (import) or lysate (integration) were incubated with
translation product and ATP for 15 min in the light at 25 C. Chloroplasts were
recovered and lysed. Recovered membranes were washed, solubilized with 1% digitonin,
and processed for BN PAGE. The approximate sizes of complexes are given to the left
of the fluorograph. The approximations are based on visual inspection of bands on the
blue native gel and photosystem complex sizes. (B) Chloroplasts were incubated with
translation product and ATP for 30 min in the light at 25 C. Chloroplasts were
recovered and lysed. Recovered membranes were washed, solubilized with 0.5%
digitonin, and processed for BN PAGE. Positions of molecular weight markers are
indicated to the left of the panel.











Import Integration ,


SC< 0 C 0 '> C 0C 0 '
/ '(N 0/ 0 00 /


880 kDa -

440 kDa -

132 kDa -


66 kDa -


Lane: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21


Figure 4-4. Incorporation of in vitro-translated components into native complexes.
(continued). (C) Chloroplasts (Import) were incubated with ATP and translated
precursors pcpTatC, pHcfl06, and pTha4 for 15 min in the light at 25 C. Lysate
(Integration) was incubated with ATP and translated mature proteins for 15 min in the
light at 25 C. Chloroplasts were recovered and lysed. Recovered membranes from
import or integration assays were washed, solubilized with 1% digitonin, and processed
for BN PAGE. Positions of molecular weight markers are indicated to the left of the
panel. A portion of the sample in lane 5 was diluted with three volumes solubilized
membranes in BN sample buffer and loaded in lane 6. Lanes labeled TP were loaded
with translation product in BN sample buffer. Lanes labeled TP + Membr were loaded
with translation product and solubilized membranes in BN sample buffer. Lanes labeled
Integrated exhibit samples of solubilized membranes from integration assays with lysate.
Lanes 5, 7, 8, and 9 are equivalent to lanes 14, 20, 15, and 21 respectively.



similar results (Fig. 4-4C, lanes 9 and 21). Again, species migrating at the same

molecular weights could be generated by adding translation product to solubilized

membranes (Fig. 4-4C, lanes 18 and 19). Therefore, the physiological relevance of

complexes of Tha4 is questionable.

Integration of In Vitro-translated Components Influenced Efficiency of Subsequent
ApH-dependent Pathway Transport.

Because in vitro-translated mHcfl06 and mTha4 associated with membranes in

amounts comparable to endogenous components, it seemed reasonable that integration of










native or mutated forms of the components might influence translocation of ApH-

dependent pathway substrates. In the experiment displayed in Fig. 4-5A, mHcfl06 or

mHcfl06 E/Q was integrated into unwashed-membranes. The membranes were

recovered and then incubated with iOE23 under transport permissive conditions.


S Membranes
"4' Membranes g>


Membranes




mHcfl06 &
iOE23 -

mOE23 -


4W2 Yan e.m -.


Labeled mHcfl06
Unlabeled mHcfl06

Labeled mHcfl06 E/Q
Protease


+ + +


+ + -


+ + +
+ +


Lanes: 1 2 3 4 5 6 7 8 9 10 11



Figure 4-5. Integration of in vitro-translated components influenced efficiency of
subsequent ApH-dependent pathway transport. Substrate iOE23 was generated from
mRNA and wheat germ extract, diluted four-fold, and adjusted to IB and 30 mM
unlabeled leucine. All component-substrates were generated from wheat germ TnT and
diluted as described in Material and Methods. Chloroplasts were lysed and unwashed
membranes were adjusted to 2 mg/ml chlorophyll. Assays were conducted by incubating
13 ptl of membranes and 31 ptl of component-substrate for 10 min in the light at 250C. An
additional 31 ptl of component-substrate was added to the assay and incubation continued
10 min. Control assays (lanes 2 and 3 of both panels) received sequential 31 ptl aliquots
of equivalently diluted wheat germ TnT. Membranes were recovered by centrifugation
and washed with IBM. Membranes were resuspended in 50 p1l IBM, 25 ptl iOE23 was
added, and the assays were returned to 250C and light for 10 min. Membranes were
recovered, protease treated where indicated, and processed for SDS-PAGE and
fluorography. (A) Labeled component-substrate mHcfl06 was integrated in lanes 5 and
6. Unlabeled component-substrate mHcfl06 was integrated in lanes 7 and 8. Labeled
component-substrate mHcfl06 E/Q was integrated in lanes 10 and 11.













B Membranes / Membranes / Membranes
-------- ---------- 4(s --------
iOE23 ..

mOE23 -



mTha4 Ao 4




Labeled mTha4 + + +
Labeled mTha4 E/Q + + +
Protease + + +

Lanes: 1 2 3 4 5 6 7 8 9




Figure 4-5. Integration of in vitro-translated components influenced efficiency of
subsequent ApH-dependent pathway transport (continued).. (B) Labeled
component-substrate mTha4 was integrated in lanes 5 and 6. Labeled component-
substrate mTha4 E/Q was integrated in lanes 8 and 9.



Thylakoids having previously integrated radiolabeled (lanes 5 and 6) or unlabeled (lanes

7 and 8) mHcfl06 accumulated more mature OE23 (mOE23) than control thylakoids

(lanes 2 and 3). Thylakoids having previously integrated mHcfl06 E/Q accumulated

slightly less mOE23 (lanes 10 and 11) than control thylakoids. A similar experiment

conducted on the same day comparing thylakoids previously incubated with mTha4 or

mTha4 E/Q revealed little difference in mOE23 accumulation among assays (Fig. 4-5B).

The experiments displayed in Fig. 4-5 were repeated numerous times. The results

were extremely variable. Stimulation of mOE23 accumulation following mHcfl06









integration was frequently, but not always, seen and when seen was less pronounced than

exhibited in Fig. 4-5. Likewise, inhibition of accumulation of mOE23 by previously

integrated mHcfl06 E/Q was rarely seen and never again as evident as in Fig. 4-5. The

reason for variability in results is unknown. Because the experiment was well controlled

internally and because less pronounced effects (but having the same trends) were

occasionally observed, I believe the variability resulted from some uncontrolled

physiological factor. The decreased mOE23 accumulation following mHcfl06 E/Q

integration would be expected if mHcfl06 E/Q were displacing endogenous mHcfl06 in

the 700-kDa complex. However, the results from the experiment in Fig. 4-4C did not

demonstrate an association of mHcfl06 E/Q in the 700-kDa complex. It is possible that a

transient interaction is occurring which is not being captured by the BN PAGE

methodology.

Investigation of the Interaction of In Vitro-translated Components with Maize
Membranes.

The maize mutant hcfl06 results in pale green non-photosynthetic seedlings that

die after expansion to three or four leaves (Voelker and Barkan, 1995). Chloroplasts

isolated from hcf]06 are devoid of the Hcfl06 protein, but do contain wild-type amounts

of Tha4 (Walker et al., 1999). Bacterial TatC was demonstrated in a pulse-chase

experiment to be unstable in the absence of the Hcfl06 orthologue TatB (Sargent et al.,

1998). It is not known whether hcf]06 chloroplasts have endogenous cpTatC.

Maize hcf]06 chloroplasts are unable to export ApH-dependent pathway

precursors to the thylakoid lumen (Settles et al., 1997). A previous attempt to restore

transport competence in mutant chloroplasts by introducing Hcfl06 protein was

unsuccessful (Cline, unpublished). Although Hcfl06 was associated with mutant









membranes, the recovered membranes were unable to transport pOE17. One possible

reason transport competence was not restored is that the hcfl06 thylakoids lacked

endogenous cpTatC. I attempted to restore transport competence by integrating both

Hcfl06 and cpTatC.

Maize cpTatC has not been cloned. Therefore, it was necessary to attempt

reconstitution of ApH-dependent pathway machinery using cpTatC from pea (ps cpTatC).

Hcfl06 is available from maize and pea in its precursor (zm pHcfl06 and ps pHcfl06)

and its mature (zm mHcfl06 and ps mHcfl06) forms. It was unknown how well pea

components would integrate into maize membranes and whether pea and maize

components would be able to associate in the 700-kDa complex.

Import assays using intact organelles and transport assays using lysate were

conducted on chloroplasts isolated from wild-type (cv. Truckers Favorite) maize.

Membranes were recovered and prepared for BN PAGE. Aliquots of solubilized

membranes were diluted in SDS loading buffer. Fig. 4-6A is a fluorograph displaying the

results from SDS-PAGE of those aliquots. Integration of both ps mHcfl06 (lane 4) and

zm mHcfl06 (lane 8) was efficient. However, import of zm pHcfl06 was comparatively

inefficient (lane 6) and ps pHcfl06 was imported minimally (lane 2). Import ofps

pcpTatC was efficient (lane 10), but there was no apparent integration of ps mcpTatC

(lane 12).

Analysis of samples by BN PAGE revealed that integrated pea components were

able to interact with endogenous maize components in the 700-kDa complex (Fig. 4-6B).

Because of the anticipated efficiency of integration, samples from assays using mHcfl06

were diluted 1:6 with solubilized membranes prior to loading them onto the gel. The











ps pHcf106 ps mHcf106


zm pHcf106 zm mHcf106 ps pcpTatC ps mcpTatC


TP sol TP sol TP sol TP sol TP sol TP sol






Lane: 1 2 3 4 5 6 7 8 9 10 11 12


C' C



880 kDa

440 kDa


Lane: 2 4 6 8 10 12


Figure 4-6. In vitro-translated ApH-dependent pathway components derived from
both pea and maize integrated into maize membranes and associated with the
endogenous 700-kDa complex. Component-substrates ps pHcfl06, ps pcpTatC, and ps
mcpTatC were generated from wheat germ TnT. Other component-substrates were
generated from mRNA and wheat germ. Intact chloroplasts were incubated with ATP
and component-substrates having transit peptides for 30 min in the light at 250C. Lysate
was incubated with ATP and mature component-substrates for 10 min in the light at
25C. Chloroplasts were recovered and lysed. Recovered membranes were solubilized
with 0.5% digitonin and processed for analysis by BN PAGE. (A) Each pair of lanes is
labeled above with the identity of the component-substrate integrated. Aliquots from the
solubilized membranes (sol) were diluted with SDS loading buffer and analyzed by SDS-
PAGE (12.5% acrylamide) and fluorography. TP is translation product as indicated
above the lanes. (B) Samples were analyzed by BN PAGE and fluorography. Lanes are
numbered to correspond to samples in (A). Samples in lanes 4 and 8 were diluted with 5
volumes solubilized membranes in BN sample buffer. The identity of the integrated
component-substrate is indicated above each lane. Positions of molecular weight markers
are indicated to the right of the panel.


A









interaction of ps cpTatC with endogenous maize Hcfl06 implies that imported ps

pcpTatC would be capable of interacting with exogenous (in vitro-translated and

integrated) Hcfl06 from either pea or maize.

As a first step in reconstitution of transport competence to maize hcfl06 mutants,

I attempted to find conditions under which I could achieve interaction of in vitro-

translated (exogenous) Hcfl06 with cpTatC (either exogenous, or endogenous). If

interaction could be achieved, I reasoned that I would be able to visualize it as the 700-

kDa complex on a fluorograph following BN PAGE. Because zm pHcfl06 had imported

significantly better than ps pHcfl06 into wild-type maize chloroplasts, I chose zm

pHcfl06 and zm mHcfl06 for these assays. All assays were run in parallel with both

wild-type and mutant chloroplasts. Samples from mutant chloroplasts are indicated in

Fig. 4-7 by asterisks at the top of the lanes. The integration of components was attempted

separately, simultaneously, and sequentially. Samples were recovered and prepared for

BN PAGE. Aliquots of solubilized membranes were diluted in SDS loading buffer. Fig.

4-7A is a fluorograph displaying the results from SDS-PAGE of those aliquots. Fig. 4-

7B is a fluorograph of samples analyzed by BN PAGE.

Using intact chloroplasts, zm pHcfl06 was imported and processed by both wild-

type and mutant organelles (Fig. 4-7A, lanes 2 and 3). However, only wild-type

chloroplasts were able to integrate Hcfl06 into the 700-kDa complex (Fig. 4-7B, lane 2

versus lane 3). Import and processing ofps pcpTatC was inefficient in wild-type

chloroplasts (Fig. 4-7A, lane 5) and was not detectable in mutant chloroplasts (Fig. 4-7A,

lane 6). Wild-type chloroplasts were able to integrate cpTatC into the 700-kDa complex











A


* TP *


simultaneous
*


sequential
* *


+ + -


+ + + + +


4 5


+ +


+ + + + +
6 7 8 9 10 11 12 13


c* *


c* *


880 kDa -

440 kDa -


Lane: 2 3 5 6 7
Exposed 42 days


Lane: 10 11 12 13
Exposed 3 days


Figure 4-7. Maize mutant hcfl06 chloroplasts did not incorporate in vitro-
translated component-substrates into a 700-kDa complex. Lanes marked indicate
membranes derived from hcf]06 mutant chloroplasts. Component-substrate ps pcpTatC
was generated from wheat germ TnT. Other component-substrates were generated from
mRNA and wheat germ. Lanes 1 through 3 and 5 through 8 exhibit samples from assays
conducted by incubating intact chloroplasts with ATP and the precursors indicated for 10
min in the light at 250C. Lanes 10 through 13 exhibit samples from assays conducted by
incubating intact chloroplasts with pcpTatC or buffer for 10 min in the light at 250C.
Chloroplasts were recovered without repurification and lysed. The lysate from assays
exhibited in lanes 10 through 13 was subsequently incubated with zm mHcfl06 for 10
min in the light at 250C. In all samples, recovered membranes were solubilized with
0.5% digitonin, and processed for BN PAGE. (A) Aliquots of samples were diluted with
SDS sample buffer and analyzed by SDS-PAGE (12.5% acrylamide) and fluorography.
TP indicates translation product. (B) Lanes are numbered to correspond to assays in (A).
Samples were analyzed by BN PAGE and fluorography. Positions of molecular weight
markers are indicated to the left of each panel. The time of exposure of film is indicated
below each panel.


zm pHcfl06
ps pcpTatC
zm mHcfl06
Lane:


+ + + -


1 2 3


880 kDa -

440 kDa -









(Fig. 4-7B, lane 5), but no integration was detectable in mutant chloroplasts (Fig. 4-7B,

lane 6).

Incubation of zm mHcfl06 with lysate from either wild-type or mutant

chloroplasts resulted in membrane association (Fig. 4-7A, lanes 12 and 13). However,

only wild-type lysate was able to integrate mHcfl06 in the 700-kDa complex (Fig. 4-7B,

lane 12 versus lane 13).

Because imported cpTatC may be rapidly degraded in the absence of endogenous

Hcfl06, I simultaneously imported both components. The sizes of mcpTatC and

mHcfl06 are too close for bands generated on the fluorograph of a 12.5% SDS gel to be

resolved. Membrane association of processed components in wild-type chloroplasts was

detected in Fig. 4-7A, lane 7 and in mutant chloroplasts in Fig. 4-7A, lane 8. Only wild-

type chloroplasts were able to integrate components in the 700-kDa complex (Fig. 4-7B,

lane 7 versus lane 8).

Finally, because mHcfl06 is so much more efficiently integrated in lysate than

imported Hcfl06, I imported ps pcpTatC, lysed the chloroplasts, and then incubated the

lysate with zm mHcfl06. A band of mature components was visible from both wild-type

and mutant membranes (Fig. 4-7A, lanes 10 and 11). However, only wild-type

membranes incorporated components in the 700-kDa complex (Fig. 4-7B, lane 10 versus

lane 11).


Discussion

I have demonstrated that in vitro-translated cpTatC and Hcfl06 are integrated

into a 700-kDa complex that migrates in BN PAGE identically to the antibody-decorated

endogenous complex. When Hcfl06 was mutated, changing its conserved glutamate to









glutamine, it failed to integrate in the 700-kDa complex. These data imply that the

glutamate in Hcfl06 is required for interaction with cpTatC.

I have shown that it is possible to integrate quantities of in vitro-translated Hcfl06

and Tha4 comparable to the amount of endogenous proteins present in pea membranes.

The ability to track the 700-kDa complex with in vitro-translated and radiolabeled

components will provide a tool for additional biochemical studies such as binding assays.

The effects of changing other amino acids in Hcfl06 or cpTatC may further inform

investigators as to the nature of the complex interaction and translocation mechanism.

In some assays, I saw an influence of exogenously added components on the

translocation efficiency of a ApH-dependent pathway substrate. The results were highly

variable and I was unable to ascertain the reason for the variability. It is likely that some

physiological variable such as growth-room temperature was involved. However, the

experiments in which I did see the influence were well controlled internally, so that I am

convinced the effects were real. The decreased accumulation of OE23 following

integration of mHcfl06 E/Q implied that Hcfl06 E/Q had some limited interaction with

other components (presumably cpTatC) disrupting complex formation.

Although my attempt to reconstitute the 700-kDa complex in maize mutant

membranes was unsuccessful, additional experiments may reveal conditions under which

that reconstitution can be accomplished. In particular, the eventual cloning of maize

cpTatC may be key. The variety of maize from which hcfl06 is derived grew more

slowly than the Trucker's Favorite variety and generally appeared less robust. Import of

ps pcpTatC was minimal in Hcfl06 control plants. Although assays were conducted

under conditions that would have allowed integration via any of the known pathways,









integration of mcpTatC in lysate from all chloroplasts, pea or maize, was inefficient.

Certainly, lysate does not fully mimic conditions inside the intact chloroplast. One of the

most obvious differences between transport in lysate and in organello translocation is the

concentration of stromal proteins. It is unknown whether any of those proteins or other

cofactors may participate in integration of cpTatC. Given the highly efficient integration

of mHcfl06 in isolated membranes, finding conditions under which mcpTatC could

integrate might significantly improve the chances of reconstituting transport of ApH-

dependent pathway substrates in hcf]06-derived membranes.














CHAPTER 5
SUMMARY AND CONCLUSIONS



The ApH-dependent pathway of chloroplasts is unique among export systems

because of its energetic requirements, its use of a recognizable motif in substrate

presequences and because of its ability to translocate a spectrum of protein substrates.

The exclusive dependence of the pathway on the trans-thylakoid pH gradient has been

well established. Multiple studies from several laboratories have attempted to define the

presequence requirements of the thylakoid pathway and its bacterial counterpart. The

flexibility of the system has been explored, as its ability to translocate folded domains,

domains with disrupted folding, and transmembrane domains has been documented. In

spite of a decade of investigation, the mechanism by which the system moves proteins

across or into the thylakoid membrane remains an enigma.

I have taken a first look at protein translocation initiation. I have shown that the

amino-terminal segment of a translocation substrate remains on the cis side of the

membrane following transport by the ApH-dependent pathway, just as it does following

transport by other export systems. This phenomenon may reflect an evolutionary

relationship among export pathways including spontaneous protein translocation, the

prokaryotic Sec pathway, and the endoplasmic reticulum secretion system.

I have arrested the process of translocation across the thylakoid membrane and

captured a membrane-spanning intermediate. The existence of an intermediate in the

translocation event argues for a dynamic process in which the progress of export is









monitored and nurtured. Manipulation of the intermediate may allow investigation of

that process and of the components involved in it.

A model for a possible mechanism of initiation, transport and arrest of transport is

displayed in Fig. 5-1. From the unpublished data of Cline and Mori, it is known that

ApH-dependent pathway precursors can bind to the endogenous 700-kDa complex even

in the absence of a delta pH and that binding is dependent on the conserved arginines and

hydrophobic domain of the signal sequence. I propose that the perturbation of the bilayer

by the hydrophobic domain of the signal peptide in the presence of a trans-thylakoid pH

gradient results in localized proton flow. One consequence of that proton flow is a

change in the conformation of a component of the 700-kDa complex. I suggest cpTatC

as the altered component, because bacterial TatC is in association with the Tha4

orthologue, TatA (Bolhuis et al., 2001). The alteration in the configuration of cpTatC

instigates oligomerization of Tha4 forming a channel de novo at the site of translocation.

The size of the channel is controlled by the three dimensional form of the translocation

substrate in contact with the amphipathic helices of multiple Tha4 monomers. The

substrate moves through the channel spontaneously with the size of the channel

decreasing as the structure passes through. Without the spontaneous insertion of the

hydrophobic domain of the signal peptide to promote proton flow and initiate further

oligomerization, the channel is unable to adapt to a large folded domain that follows an

unfolded domain. Translocation is arrested. The channel collapses, leaving the

membrane-spanning intermediate.

By documenting the integration of in vitro-translated ApH-dependent pathway

components in endogenous complexes, I have created a tool for investigating the nature









pOE17
p17-protA
STD LTD


Protein A

linker


i17-protA


.a,


ml7-protA



w ^Protease


-=E


ApH


Fragment


Figure 5-1. Model of protein translocation initiation and arrest on the ApH-
dependent pathway. Icons representing substrate domains and channel components are
not to scale. The stromal targeting domain (STD) is cleaved in the chloroplast stroma.
The lumenal targeting domain (LTD) interacts with the endogenous 700-kDa complex
(pink hexagon) and the lipid bilayer (green lines). In the presence of a trans-thylakoid pH
gradient (ApH), the channel (brown hexagon) is formed by the oligomerization of Tha4
monomers. The size of the channel decreases during transport of the linker region.
Translocation is arrested when the protein A moiety is unable to enter the channel.
Treatment of recovered membranes with protease leads to degradation of the protein A
moiety and retention of one or more fragments inside the thylakoid lumen.


of the interaction of those components. I have demonstrated that changing one conserved

amino acid in the transmembrane domain of Hcfl06 can abrogate that component's

ability to participate in endogenous complex formation. A range of in vitro mutations

and interactions can be envisioned using similar techniques. The incorporation of

exogenously derived components may eventually allow restoration of ApH-dependent


WP~






91


pathway transport to mutant plants thereby defining component requirements and

interactions.

The biochemical investigation of the ApH-dependent pathway is beginning to

define the roles of machinery components and reveal the translocation mechanism.

Research has demonstrated both the unique nature of the pathway and its evolutionary

relationship to other export pathways.




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs