Identification and characterization of cDNA clones for chromoplast-associated proteins in tomato fruit


Material Information

Identification and characterization of cDNA clones for chromoplast-associated proteins in tomato fruit
Physical Description:
vi, 153 leaves : ill. ; 29 cm.
Lawrence, Susan D
Publication Date:


Subjects / Keywords:
Tomatoes -- Genetics   ( lcsh )
Tomatoes -- Ripening   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1993.
Includes bibliographical references (leaves 136-152).
Statement of Responsibility:
by Susan D. Lawrence.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001952611
notis - AKC9183
oclc - 31280128
System ID:

Full Text








I wish to thank Drs. Donald Grierson and Wolfgang Schuch for providing us

with clones from their ripening tomato fruit cDNA library. I would also like to take

this opportunity to thank the members of my committee Drs. Gloria Moore, Ken

Cline, Eduardo Vallejos, Don Huber and Bill Gurley for their advice and support.

I am especially indebted to Eduardo Vallejos for extensive years of committee

service. This dissertation is dedicated to Mike McCaffery for his tireless computing

assistance and to Ralph Henry for his photographic skills. Finally, I would like to

express my appreciation to all my Fifield Hall lab pals, past and present, for making

this an experience I will fondly recollect.


ACKNOWLEDGMENTS ...................................... ii

ABSTRACT ................................................ v


1 INTRODUCTION ................................. 1
2 LITERATURE REVIEW ............................ 3
Morphological Changes .............................. 5
Biochemistry and Molecular Biology ..................... 6
Changes in Plastid Structure .......................... 15
Regulation of Photosynthetic Proteins ................... 16
The Role of Ethylene in Tomato Fruit Ripening ........... 18
Transcriptional Regulation of Fruit Ripening .............. 19
Processes Associated with Chromoplast
Development ..................................... 20
Heat Shock Proteins and Oxidative
Stress .......................................... 40
Conclusions ...................................... 51

IN TOMATO FRUIT .............................. 54
Introduction ..................................... 54
Materials and Methods .............................. 55
Results and Discussion .............................. 58


Introduction ..................................... 74
Materials and Methods .............................. 75
R results ......................................... 82
D discussion ...................................... 105

5 CONCLUSIONS ................................. 118

pTOM 41 AND Ptom 111 PLASMIDS IN TOBACCO ..... 120

REFERENCES ............................................ 136

BIOGRAPHICAL SKETCH .................................. 153

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



Susan D. Lawrence

December 1993

Chairman: Dr. Gloria A. Moore
Major Program: Plant Molecular and Cellular Biology

The chloroplast to chromoplast transition during tomato fruit ripening is

characterized by a dramatic change in plastid structure and function. The thylakoid

membranes of photosynthetically competent chloroplasts are broken down and the

carotenoid pigment lycopene is deposited within the developing chromoplast. A goal

of this investigation was to determine whether this process is mediated by an increase

in the steady-state level of RNA for plastid-targeted proteins. Assays for import of

radiolabeled translation products into isolated pea chloroplasts were used to monitor

levels of chromoplast-targeted proteins at four stages of tomato fruit development.

During development, striking increases in levels of translatable RNA for two such

proteins were found. Additionally, the import of in vitro translation products was

examined for seven individual cDNA clones known to encode RNA that increases

during fruit ripening. Three of these clones produced in vitro translation products

that imported into pea chloroplasts. One clone, pTOM5, has recently been shown by

others to encode the enzyme responsible for the first committed step in the synthesis

of lycopene.

Subsequent sequence analysis of another cDNA, pTOM 111, revealed that it

encodes a plastid-localized low MW heat shock protein. Not only does this transcript

increase during development at normal temperatures, but it is increasingly inducible

by elevated temperatures as ripening progresses. This is the first report of the

developmental induction of a plastid-localized low MW heat shock protein.

Identification of the induction of a heat stress protein during this transition suggests

that either the plastid is subjected to an increasingly stressful environment or, as is

consistent for the developmental induction of other heat stress proteins, a massive

structural reprogramming is underway.

These studies imply that there is synthesis and import of new proteins during

the transition from chloroplast to chromoplast, and that the plastid conversion is an

active developmental program rather than a simple decline in synthesis of the

photosynthetic apparatus. Furthermore, these results demonstrate the utility of this

method for identification of structural genes involved in plastid morphogenesis.


Ripening of tomato fruit involves a wide variety of developmental changes

that include but are not limited to changes in cell wall structure, in metabolism, in

ultrastructure, and an especially dramatic change in the form and function of the

plastid. Biochemical and molecular studies suggest that most or all of these changes

are mediated by the enhanced expression of ripening specific genes. I have chosen

this system as a means to investigate plastid development and interconversion.

We are interested in determining how the transition from chloroplast to

chromoplast is regulated, e.g., does a subset of plastid proteins increase during this

process? In an initial examination of this question, we have asked whether an

increase in the accumulation of RNA for a subset of plastid proteins correlates with

the chloroplast to chromoplast transition. Since transcription and/or translation of

the plastid genome seems to diminish during chromoplast formation (Kobayashi,

1991), chromoplast-specific proteins are probably encoded in the nucleus. Nuclear-

encoded plastid proteins are synthesized in the cytosol and posttranslationally

imported into plastids. The import of proteins into plastids is a highly specific

process (Keegstra, 1989). The proteins are made as precursors with transient amino

terminal "transit peptides," which are both necessary and sufficient for the correct

targeting into this organelle. It is this latter observation that makes the import into

plastids a potentially diagnostic assay for the identification of chromoplast associated


The approach described within this dissertation has been to monitor

chloroplast import of in vitro translation products from RNA of tomato fruit isolated

at different ripening stages. We have also studied translation products programmed

with transcripts from cDNAs (Slater et al., 1985) whose steady state RNA level

increases during ripening. Since individual cDNAs were analyzed, subsequent

sequence analysis allowed further characterization of the proteins that mediate this

transition. Eventually the role of chromoplast-associated proteins in chromoplast

development can be assessed directly via antisense technology. The final outcome of

such a study should be an understanding of the role new protein synthesis plays in

this process and the isolation, identification and characterization of genes that

mediate this transition.


The manner in which the transition from chloroplast to chromoplast is

regulated is largely unknown. Electron micrographic analysis has shown that a

massive reordering of plastid structure occurs during the conversion (Thelander et

al., 1986). In tomato fruit, this is characterized by the overall breakdown of

chloroplast specific components, e.g. the thylakoid membrane system and starch, with

the concomitant accumulation of membrane-associated deposits of the carotenoid

lycopene (Thelander et al., 1986). It has not been determined whether the

breakdown of chloroplast structures is caused simply by the decrease in synthesis of

photosynthetic proteins or if increases in degradative enzymes actually enhance this

process. An examination of nuclear-encoded, plastid-targeted, photosynthetic proteins

has shown that their transcription stops 5 to 10 days before the fruit reaches full size

(Piechulla et al., 1986). Analysis of plastid-encoded transcripts has demonstrated that

some are greatly reduced, whereas others persist, and one appears to increase during

chromoplast formation (Kobayashi, 1991; Richards et al., 1991). Similarly, little has

been ascertained about the function or expression of genes involved in the synthesis

of chromoplast structures as few clones specific to this developmental stage have

been available for analysis (for example, Kuntz et al., 1992).

In considering what aspects of fruit ripening directly affect the proteins that

are targeted to the plastid, it quickly became apparent that two types of functions

could cause increases in proteins directed to the developing chromoplast. First,

proteins involved in the degradation of the photosynthetic apparatus or the build up

in carotenoids are excellent candidates for chromoplast-targeting and enhanced

synthesis during fruit ripening. A broad description of plastid senescence will be

considered, since that transition also involves the degradation of chloroplast

structures and because an examination of RNA from senescing leaves has shown that

many of the same transcripts are also present. Secondly, other aspects of maturation

may have indirect effects on the environment in which the developing plastid resides.

For example, the autocatalytic production of ethylene, the respiratory climacteric or

an increase in oxidative stress during ripening could secondarily create increases in

plastid-targeted proteins. Since the plastid is responsible for a number of basic

biosynthetic functions, such as the synthesis of aromatic amino acids, fatty acids and

8-aminolevulinic acid these changes outside of the plastid could create a need for

substrates synthesized within this organelle.

A consequence of the work described in this dissertation was the

characterization of a cDNA that encodes a plastid-localized low molecular weight

(MW) heat shock protein (hsp). Because transcript for this hsp increases during

ripening at normal temperatures, it was necessary to assess how this unexpected

finding fit into chromoplast development or fruit ripening. Since oxidative stress has

been associated with senescence and hsps have been induced by oxidative stress in

animal tissues, a consideration of heat stress proteins and oxidative stress during fruit

ripening will also be examined.

Morphological Changes

The overall changes that occur during the ripening of tomato fruit are as

familiar to ones mother as they are to any post-harvest physiologist. Full sized green

fruit harvested from the garden and brought into the kitchen will ripen normally. The

color changes observed during the ripening of tomato fruit are due to the breakdown

of chlorophyll and the deposition of lycopene within the fruit plastids. The breaker

stage is identified by the appearance of external color at the blossom end. The exact

fruit color that then develops is dependent on variety and consequently, the fruit may

turn yellow, pink, or orange-red. As the fruit becomes uniformly red, the cell walls

begin to soften. The characteristic flavor of ripe tomatoes is created by a

combination of glucose, fructose, citric and glutamic acids, which accumulate in the

vacuoles, along with the synthesis of hundreds of volatile compounds (Davies and

Hobson 1981; Grierson, 1986).

The young fruit is green because it contains chlorophyll in photosynthetically

functional chloroplasts. It is considered immature green (IMG) until it reaches its

full size. Growth is complete at mature green (MG). A green tomato may appear

morphologically quiescent; however, it is undergoing important ripening changes. Its

locular tissue transforms from a solid to a gelatinous substance, it can respond

rapidly to exogenous ethylene and the endogenous ethylene level also increases at

this time. By late MG, carotenoids begin to appear in the locules. Thus the color

change begins in the interior of the fruit (Kader and Morris, 1976). After the fruit

has passed through mature green, the transformation to a red ripe soft tomato

proceeds rapidly.

Biochemistry and Molecular Biology

Although the change in fruit color is a convenient method of classifying

tomato fruit ripening, a number of other biochemical changes also occur. For

example, a dramatic rise in respiration and ethylene production takes place. This

phenomenon is known as the climacteric. Climacteric ethylene levels reach about one

half of their maximum by the breaker stage (Picton and Grierson, 1988). The

respiratory peak occurs soon after the rise in ethylene (Grierson, 1986).

Alterations in Basic Metabolism During Fruit Ripening

Changes take place in the type of carbohydrate that accumulates as tomato

fruit ripens. Starch granules transiently appear between 14-28 days postanthesis in

the immature green chloroplasts (Davies and Cocking, 1965). Although these

plastids can fix carbon at up to 35% of the rate attained by leaf plastids, the starch

granules are mostly the product of sucrose imported from the leaves (Thelander et

al., 1986). Since the import of carbohydrate from the leaves stops by MG (Thelander

et al., 1986), fruit at this stage will develop normally if harvested from the plant.

About 28-35 days after anthesis degradation of starch occurs as an increase in soluble

sugars takes place. Consequently, little starch remains in the mature green fruit

(Yelle et al., 1988). The rapid degradation of starch precedes the climacteric rise in

respiration (Davies and Cocking, 1965) and this carbohydrate is stored as glucose

and fructose in the vacuole (Grierson, 1986). The degradation of starch seems to

occur via phosphorylase (Robinson et al., 1988), and the further steps of hexose to

triose occur via the pentose phosphate pathway and glycolysis (Thelander et al.,

1986). Therefore, as the tomato fruit matures, a shift of stored carbohydrate occurs

from starch in the plastid to hexose sugars in the vacuole. In fact, the red ripe

tomato contains no measurable starch and 47% hexose by dry weight (Davies and

Hobson, 1981). The changes in stored carbohydrate precede the breakdown of the

thylakoid membranes and consequently, any plastid associated enzymes responsible

for starch degradation are not expected to increase during chromoplast development.

It is unclear what causes the increase in CO, during ripening and,

consequently, the molecular regulation of this increase is also unknown (Solomos,

1988). Fluctuations in enzymes involved in the tricarboxylic acid (TCA) cycle and

glycolysis occur but it has not been determined whether these changes are due to a

variation in the level of enzymatic effectors or are caused by alterations in the

amount of de novo synthesis of these enzymes during ripening. Solomos (1988)

reviewed the literature on respiration in ripening fruit and concluded that the

respiratory burst measured is probably a result of an increase in substrate level

phosphorylation caused by enhanced glycolysis. An increase in the activities of

cytosolic NADP linked malic enzyme (Goodenough et al., 1986) and the glycolytic

enzyme phosphofructokinase (PFK) have been observed (Grierson, 1986). PFK is a

control point in glycolysis and an increase in this enzyme activity probably results in

an increase in carbon flow through this pathway (Solomos, 1988). Indeed, an increase

in the accumulation of ATP is generally noted during the respiratory rise (Solomos,

1988). Investigation of mitochondrial electron transport in ripening fruit suggested

that it remains unchanged (Solomos, 1988). However, the TCA cycle enzymes citrate

synthase, malate dehydrogenase, NAD-linked malic enzyme, and isocitrate

dehydrogenase decrease in ripening tomato fruit (Grierson, 1986; Jeffery et al.,1986).

Overall, these data do not explain why an increase in respiration is measured during

ripening. Since these changes are underway during the plastid conversion while

photosynthesis is declining, it is unclear how this may effect the synthesis of proteins

targeted to the plastid.

Other Changes Associated with Fruit Ripening

Apparently, the level of the anoxia-responsive enzyme, alcohol dehydrogenase

(ADH), increases during ripening in tomato fruit (Longhurst et al., 1990; Van Der

Straeten et al., 1991). Longhurst et al. (1990) concluded that the increase in ADH

during ripening may be caused by the cytoplasmic increase in pH. It has been

hypothesized that low oxygen levels could result within such a bulky organ, but by

exposing ripening fruit to different levels of 02 and CO2 these authors surmised that

the change in pH leads to an induction of this enzyme.

As fruit matures, the cell walls soften and an increase in soluble pectin occurs

(Grierson, 1986). Characterization of the regulation of the enzymes involved in cell

wall degradation has begun. These data suggest that the pattern of expression for

enzymes involved in cell wall degradation is complex and can even involve more than

one isozyme (DellaPenna et al., 1986; Grierson et al., 1986; Fisher and Bennett,

1991; Harriman et al., 1991). For example, antisense mutagenesis of the cell wall

softening enzyme, polygalacturonase (PG), has shown that although this enzyme is

dramatically induced during fruit ripening, loss of this enzymatic function has little

overall effect on maturation (Sheehy et al., 1988; Smith et al., 1988 and 1990).

The synthesis of hundreds of volatile compounds occurs as tomato fruit ripens

(Davies and Hobson, 1981). However, little is known about how these compounds

are synthesized during maturation.

Oxidative Stress and Fruit Ripening

Since a chromoplast-associated low MW heat shock protein was identified in

this work, it is important to understand the role this stress protein may play in fruit

development. Oxidative stress occurs when the cell's ability to respond is exceeded

by the level of reactive oxygen species present (Barja de Quiroga, 1992). The

induction of hsps by compounds that cause oxidative stress has been shown in animal

tissues (Drummond and Steinhardt, 1987; Cajone and Bernemmi-Zazzera, 1988;

Courgeon et al., 1988). In addition, oxidative stress has often been associated with

ripening and senescence. One theory postulates that a buildup in the deleterious


effects of free radicals results in the physiological manifestations of aging (Harman,


Free radical production during normal cellular metabolism

Reactive oxygen compounds can cause damage to proteins and especially

membrane components of the cell. A number of normal cellular functions result in

the production of these species, which can be a side effect of reactions that reduce

molecular oxygen. For example, at least one percent of the oxygen utilized during

respiratory electron transport is incompletely reduced (Fleming et al., 1992).

Furthermore, an increase in basal metabolism can result in an increase in the

amount of reactive oxygen generated even without invoking a breakdown in oxidative

phosphorylation (Barja de Quiroga, 1992). Photosynthesis is another normal cellular

function during which a large amount of free radicals can form (Scandalios, 1993).

In addition, some enzymatic reactions directly produce reactive molecules. Examples

of this latter group include xanthine oxidase (an ATP degrading enzyme) and beta

oxidation of fatty acids (Scandalios, 1993). Finally, certain environmental stresses can

enhance the level of reactive molecules. For example, radiation, herbicides,

pathogens, hyperoxia, and temperature fluctuations can induce free radical

production (Scandalios, 1993). Perhaps a buildup of these deleterious compounds

occurs during the ripening of tomato fruit. Since any cellular function that involves

the reduction of oxygen causes production of reactive oxygen, either autocatalytic

ethylene synthesis or the respiratory climacteric would be excellent sources of free


One source of reactive oxygen is lipoxygenase activity, which increases during

ripening (Thompson et al., 1987). This enzyme incorporates oxygen into

polyunsaturated fatty acids (Hatanaka et al., 1992) and leads to a build up of the

reactive oxygen compounds hydroperoxide or superoxide. Recently, Bowsher et al.

(1992) have purified a membrane associated lipoxygenase from breaker stage tomato

fruit. They find that the activity fractionates with the thylakoid membrane.

Consequently, this could result in increases in the amount of reactive oxygen species

in the developing chromoplast.

The change in membrane integrity during fruit ripening

A decrease in membrane fluidity, which may be created by oxidative stress,

occurs during ripening and senescence along with a progressive loss of specialized

membrane function (Thompson et al., 1987). Many aspects of membrane structure

can affect the level of membrane viscosity. For example, the length or degree of

unsaturation of the hydrocarbons and the level of sterols all can play a part.

Although a decrease in membrane fluidity occurs in the plasma membrane of tomato

fruit (Legge et al., 1986) and pepper microsomal membranes (Whitaker, 1991), no

change in plastid membrane fluidity occurs in either tomato or pepper chromplasts.

Neither fluorescent membrane probes, which directly assess the structure within

specific environments of the membrane used by Legge et al. (1986), nor

measurements of the free sterol or total sterol to phospholipid ratios determined by

Whitaker (1991) could detect a change in the developing chromoplast membrane.

Considering the increase in carotenoid biosynthesis during ripening, which is

associated with plastid envelope membranes, it is appropriate that the integrity of

this membrane persists.

Free radical scavengers and oxidative stress

Initially, an increase in free radicals results in an increase in enzymes or

antioxidants that neutralize these reactive species. For example, a family of

superoxide dismutases (SOD) acts to scavenge superoxide and create oxygen and

hydrogen peroxide. Five isozymes of this enzyme have been described in tomatoes.

Two are cytoplasmic, one is mitochondrial and two occur in chloroplasts (Perl-Treves

and Galun, 1991). Also, ascorbic acid and a-tocopherol can act as antioxidants

(Leshem, 1988). A number of enzymes are also involved in reducing the levels of

hydrogen peroxide. Catalase, peroxidase and glutathione peroxidase all act in this

manner (Leshem, 1988). Additionally, glutathione reductase is required to regenerate

glutathione lost by interaction with hydrogen peroxide (Leshem, 1988).

Consequently,the increase in these enzymes is correlated with an increase in reactive

oxygen molecules. Any decrease in these enzymes with senescence, therefore, could

contribute to degeneration of membranes and disturbance of cellular function.

Increases in basal metabolism may induce SODs in plant tissue. Bowler et al.

(1989) concluded that the mitochondrial SOD may respond to changes in the activity

of respiratory electron transport, because the induction of this enzyme in tobacco

correlated with increased activity for cytochrome oxidase.

In addition, oxidative stress in one compartment can lead to the induction of

free radical scavengers in another part of the cell. This was shown when paraquat,

which is thought to act mainly on chloroplasts, was applied to tobacco and caused

a strong induction of SOD transcript in the mitochondria, cytoplasm and chloroplast

(Tsang et al., 1991).

The level to which SOD is induced within a cellular compartment may depend

on the amount of SOD that is currently present. This was the explanation of Perl-

Treves and Galun (1991) when they found that the addition of paraquat resulted in

a greater induction of the cytoplasmic SOD in comparison to the plastid message in

tomato leaves. The level of plastidial enzyme activity in mature leaves is already high

in comparison with transcript level. They suggested that induction of plastidial SOD

transcript may therefore be unnecessary.

Ethylene has been shown to induce cytoplasmic and plastidial SOD transcripts

and enzyme activity in tomato leaves (Perl-Treves and Galun, 1991) and

mitochondrial SOD message in tobacco leaves (Bowler et al., 1989). In response to

ethylene, therefore, an increase in reactive oxygen species has been noted and

enzymes within all three cellular compartments can be affected.

Free radical scavengers in ripening fruit

The role of free radical scavengers during fruit ripening, however, is somewhat

confusing. Peroxidase has been shown to decrease during ripening (Rabinowitch et

al., 1982) or peak during the pink stage in tomato fruit (Rothan and Nicolas, 1989).

Perhaps the techniques used by Rabinowith et al. (1982) were not sensitive enough

to detect this peak of activity. The transient increase in basic peroxidase activity

paralleled the increase in ethylene (Rothan and Nicolas, 1989). Since the conversion

of 1-aminocyclopropane-l-carboxylate (ACC) to ethylene requires oxygen, this may

lead to a transient increase in reactive oxygen molecules and an increase in this free

radical scavenger.

Changes in SOD have been reported during tomato fruit ripening but results

were contradictory. SOD activity has been found to peak at the pink or orange phase

(Rabinowitch et al., 1982; Livine and Gepstein, 1988). However, no change in the

two plastidial SOD isozymes was reported by Perl-Treves and Galun (1991), nor was

a peak of the cytoplasmic or plastidial SOD message noted during ripening. In fact,

they found that the plastid-associated message actually decreased with maturation.

They explain these contradictory results by suggesting that the enzyme could be quite

stable and turnover very slowly. It is hard to explain all of the varied data reported

by these three groups. More work needs to be done before a conclusion can be made

about free radical scavengers and ripening in tomato fruit. Since the messages for the

additional plastidial and cytoplasmic SODs have not been examined, perhaps these

additional data will help reconcile these dissimilar observations.

Pepper chromoplasts do appear to respond to an oxidative stress during fruit

maturation (Romer et al., 1992). The antioxidant producing enzyme cysteine synthase

was purified and a cDNA was isolated. They found not only did this enzyme increase

during ripening but the activity of glutathione reductase, glutathione synthetase and

glutathione content within chromoplasts also increased during their development.

Since all these compounds protect organisms from oxidative damage and an increase

in a-tocopherol also occurs during ripening, they concluded that these increases in

antioxidants and peroxide scavenging compounds may allow protection of the

overaccumulated carotenoids and other membrane components from oxidative stress.

Since the results from tomato fruit on the induction of free radical scavengers

is contradictory, information from other systems may help predict what occurs in

tomato fruit chromoplasts. Leaf tissue responds to an increases in basal metabolism

and the addition of ethylene by inducing SOD (Bowler et al., 1989; Perl-Treves and

Galun, 1991). In addition, a compound expected to affect predominantly one

compartment can induce SOD in other portions of the cell (Tsang et al., 1991).

Finally, the level of induction may be dependent on current levels of enzyme activity

(Perl-Treves and Galun, 1991). Perhaps the results from pepper chromoplasts where

increases in a number of these free radical scavengers has been measured, explains

how the chromoplast, unlike the plasma membrane, avoids oxidative damage and

maintains membrane integrity (Whitaker, 1991; Romer et al., 1992). One might

predict that the plastid can respond to an environment that contains an increasing

amount of reactive oxygen and consequently it is able to maintain chromoplast

membrane integrity during ripening. These data also suggest that the production of

ethylene and/or the respiratory climacteric may induce the production of

chromoplast-associated proteins involved in free radical scavenging.

Changes in Plastid Structure

Massive changes take place in plastid ultrastructure during the chloroplast to

chromoplast transition. The thylakoid membranes are dismantled and chlorophyll

degrades (Thelander et al., 1986). Concomitantly, lycopene begins to accumulate in

the stroma. The green flesh mutant, which is inhibited in chlorophyll degradation

during ripening, is helpful in the determination of how pigment accumulates in the

developing chromoplast. Electron micrographs show that the granal stacks persist in

this mutant even as pigment is deposited (Goodwin, 1980), and suggests that

lycopene builds up around the disintegrating thylakoids. By the time a normal tomato

is fully ripe, the chromoplasts contain little else but large stores of lycopene, which

accumulate as lipophilic globules and crystals of pigment (Harris and Spurr, 1969).

The type of chromoplast that develops may be dependent on the type of

carotenoids that accumulate. The 'r' mutant, for example, with its low level of both

lycopene and B-carotene, contains only plastoglobules (Goodwin, 1980).

Regulation of Photosynthetic Proteins

Transcripts for the nuclear-encoded components of photosynthesis decrease

long before the breakdown of the thylakoid membranes. The mRNA for rbcS and

cab (Rubisco small subunit and chlorophyll a/b binding protein respectively) are not

detected after 30 days post anthesis (Piechulla et al., 1985, 1986). This loss of nuclear

transcripts occurs 5 to 10 days before MG, and precedes the end of chlorophyll

synthesis and the disassembly of the thylakoid membranes by 2-3 weeks.

The plastid encoded messages rbcL and psbA (Rubisco large subunit and the

32kD photosystem II protein), however, persist throughout development (Piechulla

et al., 1986). Most of the other plastid transcribed proteins, such as the photosystem

I and II proteins (psaA psbB psbC or psbD), all diminish in message levels

throughout development.

Once ripening has progressed to an intermediate stage of development beyond

mature green, levels of photosynthetic activity have greatly diminished and this is

reflected in decreased amounts of photosynthetic proteins. For example, by the

yellow stage of development, the amount of LHCP protein has declined to 10% of

the level present in IMG fruit (Livine and Gepstein, 1988).

At the yellow-orange stage, the level of Rubisco activity has diminished to 2%

of the level present at MG (Piechulla et al, 1987) while the amount of measurable

protein for the large subunit has also decreased. The level of photosynthetic electron

transport has decreased to 13% of what was observed at MG (Piechulla et al., 1987).

Only plastocyanin and the 32kD protein were detected by this stage of development.

These data underscore the dramatic change in the protein components of the

plastid that occurs between mature green and this intermediate stage of ripening. It

is unclear from this type of analysis whether the loss of photosynthetic proteins is

simply due to a decrease in synthesis or is also caused by an enhanced production

of degradative enzymes.

Examination of plastid gene expression during chromoplast development has

shown a reduction in the level of plastid transcript during this process. It is uncertain

whether this is the result of posttranscriptional regulation (Gruissem et al., 1989;

Marano and Carillo, 1991), or whether the rate of transcription is decreased due to

methylation of the plastid genome (Ngernprasirtsiri et al., 1988; Kobayashi et al.,


1990). The level of translation is also dramatically reduced in Capsicum plastids

(Kuntz et al., 1989) where ultrastructural analysis has shown that red fruit contain

no ribosomes or rRNA (Carde et al., 1988). Richards et al. (1991) have identified

a plastid transcript that is more abundant in chromoplast RNA than chloroplast

RNA of tomato fruit. However, given the lack of functional translation in pepper

chromoplasts and the decreased amount of transcripts present in tomato

chromoplasts, it is expected that most proteins that are increasingly present in

chromoplasts will be nuclear-encoded.

The Role of Ethylene in Tomato Fruit Ripening

The rise in ethylene during the climacteric is mediated by de novo synthesis

of at least the final two enzymes in the biosynthetic pathway (Holdsworth et al. 1987;

OIson et al., 1991; Rottmann et al., 1991). To date, only the genes for the enzymes

that catalyze these final two steps in ethylene biosynthesis have been isolated.

Ethylene is synthesized from methionine via S-adenosyl-L-methionine and 1-

aminocyclopropane-1-carboxylic acid (ACC). The enzyme that synthesizes ACC

(ACC synthase) is thought to be a pyridoxal 5'-phosphate utilizing enzyme (Yip et

al., 1990). The final step in the synthesis of ethylene is mediated by ethylene oxidase

(EFE), requires oxygen, and can be inhibited by the addition of free radical

quenchers (Matoo and Aharoni, 1988). The synthesis of ACC is considered the rate

limiting step in ethylene biosynthesis. However, EFE is also increasingly synthesized

during ripening in tomato and avocado fruit, (Holdsworth et al., 1987, McGarvey et

al., 1992) and during the senescence of carnation petals (Wang and Woodson, 1991).

Production of transgenic tomato plants that contain antisense transcripts for

the enzyme ACC synthase has shown that ripening involves both ethylene dependent

and independent pathways (Theologis, 1992). A number of laboratories have

produced transgenic tomato plants with reduced levels of ethylene (Hamilton et al.,

1990, Klee et al., 1991, Oeller et al., 1991). The most dramatic decrease occurred in

a transgenic plant carrying ten antisense copies of part of a gene for ACC synthase.

The authors reported that the plant has a 99.5% reduction in the level of ethylene

(Oeller et al., 1991) and produces fruit that remain firm and green until exogenous

ethylene is applied. If no ethylene is added, the mutant fruit still express RNA for

the ripening induced cell wall softening enzyme, polygalacturonase (PG) (Oeller et

al., 1991). This result was unexpected because ethylene is required for the expression

of PG. This suggests that the regulation of ripening by ethylene is more complex than

just simple transcriptional control. While the transcription of PG is ethylene

independent, the expression of the enzyme is dependent on the continued presence

of this hormone.

Transcriptional Regulation of Fruit Ripening

A molecular analysis of fruit ripening in tomato has led to the conclusion that

the de novo synthesis of ripening specific proteins occurs and is responsible for many

of the changes associated with maturation. Several researchers have characterized


changes in the profile of in vitro translation products that occur during development

of tomato fruit (Rattanapanone et al., 1978; Speirs et al., 1984; Biggs et al., 1986;

Lincoln et al., 1987). A number of groups have constructed cDNA libraries enriched

in ripening specific sequences (Mannsson et al., 1985; and Slater et al., 1985; Lincoln

et al., 1987). Generally, the onset of ripening in tomato fruit is mediated by an

increase in a specific subset of transcripts, including those for PG, EFE and ACC

synthase (DellaPenna et al., 1986; Grierson et al., 1986; Holdsworth et al., 1987;

Olson et al., 1991; Rottman et al., 1991).

Examination of the clones isolated by Grierson's group has shown that, for

most of their cDNAs, RNA fails to accumulate when the fruit is held at high

temperature (Picton and Grierson, 1988). High temperature (35C) has an inhibitory

effect on fruit ripening, depressing the synthesis of lycopene, PG, and ethylene

(Picton and Grierson, 1988). Perhaps this effect of high temperature represents a

heat shock response (Picton and Grierson, 1988). This was not an overall effect for

all clones, however, since RNA increased initially, for some clones, in response to

an increase in temperature. Unfortunately, many of these clones have yet to be

assigned a function.

Processes Associated with Chromoplast Development

Upon review of the work characterizing the molecular changes that occur

during the chloroplast to chromoplast transition, it is notable that mainly

photosynthetic proteins have been examined during this phase (Piechulla et al., 1985,

1986, 1987; Livine and Gepstein, 1988). To undertake the investigation of proteins

that increase during chromoplast development, a method must be found to identify

proteins specific to this process. Isolation of chromoplasts and the subsequent

comparison of proteins on SDS-PAGE from different developmental stages has been

attempted and proteins that increase in amount have been identified (Iwatsuki et al.,

1984, Bathgate et al., 1985, Wrench et al., 1987, Hadjeb et al., 1988). Using this

method, specific chromoplast-associated proteins have been isolated from only

pepper chromoplasts (Hadjeb et al., 1988). Antibody has been purified for two

proteins that increase during Capsicum chromoplast development (Newman et al.,

1989). One of these proteins binds carotenoids and is a major constituent of

Capsicum chromoplast membranes (Cervantes-Cervantes et al., 1990). However, this

protein probably binds capsanthin since the antibody does not react to any protein

present in chloroplasts or tomato fruit chromoplasts (Cervantes-Cervantes et al.,

1990). While this type of approach has resulted in the characterization of a few

proteins involved in chromoplast function, only abundant proteins can be identified

in this manner.

The plastid is involved in the synthesis of many compounds necessary for cell

function. For example, the synthesis of fatty acids, aromatic amino acids and other

substances such as 8-aminolevulinic acid have all been localized to plastids. Many of

these compounds are precursors to a wide array of secondary metabolites, so it is

possible that an increased synthesis of these plastid-synthesized compounds is

required during chromoplast development. A recent review by Hrazdina and Jensen

(1992) however documented that in many instances isozymes for the synthesis of

amino acids involved in secondary metabolism may be present in the cytoplasm

rather than in the plastids. More information on the compartmentalization of these

pathways is needed before conclusions about the localization of enzymes for

secondary metabolism can be drawn.


Enzymes responsible for the synthesis of carotenoids are obviously the most

likely candidates for proteins that may be increasingly produced during chromoplast

development. Therefore, it is necessary to consider what enzymes and cofactors may

be involved in this pathway and how it may be regulated.

Carotenoid function

In the chloroplast, carotenoids are an integral part of the photosynthetic

apparatus, where they act as accessory pigments funneling light energy into the

reaction centers. If light intensity becomes too great, the photosynthetic carotenoids

can also trap excessive light energy. Without these peripheral pigments, the plants

may photobleach (Mayfield et al., 1986). The necessity of carotenoids to forestall this

deleterious event is apparent when synthesis of carotenoids is blocked by inhibitors

or mutation.

Another function of carotenoids, however, is the pigmentation of many

flowers and fruits. Most of the orange, yellow or red pigments found in these tissues

can be attributed to carotenoids, which accumulate in the chromoplasts.

Carotenoid structure

Carotenoids are long chain hydrocarbons. The conjugated double bonds give

them their color. The synthetic pathway produces carotenes, xanthophylls and

abscissic acid. The desaturations that occur in the biosynthesis of carotenoids can

create either a cis or trans configuration. Cyclization at one or both ends can occur

(Goodwin, 1980). A beta, epsilon or gamma ring forms based on the position of the

double bond (Goodwin, 1980). The end groups can be acyclic in the case of lycopene.

If the pigment is a long chain hydrocarbon then it is a carotene. Xanthophylls on the

other hand, are carotenoids that contain oxygen.

A number of different modifications are possible to this basic long

hydrocarbon chain; 500 different structures have been described. The complement

of carotenoids found in green leaves are relatively uniform throughout the higher

plant kingdom (Goodwin and Mercer, 1983). This may be expected considering the

importance these compounds play in photosynthesis and the conservation of structure

required of the proteins and pigments of the thylakoid membranes. However, wide

structural variety occurs in carotenoids found in petals and fruit. The type of

molecule that accumulates is species dependent.

Carotenoid synthesis

The basic scheme for the synthesis of higher plant carotenoids is shown in

Figure 2-1. The production of phytoene represents the first committed step in the

synthesis of carotenoids. Usually, step-wise desaturation of phytoene leads to all-

trans-lycopene. The normal product of phytoene synthase is 15-cis-phytoene. Where


does the cis-trans isomerization take place? Beyond zeta-carotene, trans

intermediates are involved. Four stepwise removals of 2 hydrogen molecules occur

from phytoene to lycopene, as oxygen is reduced in this reaction. Proton attack of

the C1 double bond of lycopene is involved in the formation of the ring structures.

Different enzymes are responsible for the different rings that are formed. Little

interconversion of these rings occurs. Xanthophyll formation results from the

addition of hydroxyl groups.

Location of carotenoid biosynthesis

Where are the enzymes of carotenoid synthesis located? All activities after the

synthesis of isopentyl pyrophosphate (IPP) have been shown to occur in the plastids

(Gray, 1987). Compartmentalization of the enzyme activity has been best analyzed

in chloroplasts as opposed to chromoplasts, since they are most amenable to

fractionation. Mayfield (1986) examined the chloroplasts of spinach, and concluded

that phytoene desaturase is located on the inner envelope membrane. Ultrastructural

examination of developing chromoplasts suggests that pigments associate with

membrane as they are produced in the inner envelope, which is followed by

vesiculation of this membrane pigment complex (Camara and Brangeon, 1981). In

addition, examination of the lipid component of these membranes confirms that they

are a product of the inner envelope (Sitte et al., 1980).

Genetic studies have mapped all the known tomato mutants to the nuclear

genome (Tomato Genetic Cooperative Report, 1987), which suggests that the genes

involved in the synthesis of carotenoids are nuclear-encoded. Furthermore, inhibition

Isopenlenyl pyrophosplile

Geranylgeronyl pyrophosphole





v Co*C tene

y -Corolene


Fig. 2-1 Pathway for the Biosynthesis of Carotenoids in Higher Plants

of plastid translation has demonstrated that this process is not necessary for

carotenoid accumulation (Camara, 1984). Finally, the enhanced synthesis of phytoene

caused by the addition of CPTA is blocked by the cytoplasmic translation inhibitor

cycloheximide. Therefore, the evidence indicates that the enzymes required for

carotenoid synthesis are imported into the plastid.

Enzymology of carotenoid biosynthesis in tomato fruit

The purification of the carotenoid biosynthetic enzymes is difficult because all

the enzymes in this pathway beyond the synthesis of phytoene are associated with

plastid membranes. Consequently, studies that characterize the number and type of

enzymatic steps responsible for the synthesis of these compounds in higher plants

have relied on adding radioactive precursors to solubilized crude enzyme

preparations. The distribution of radioactivity between the different carotenoids is

then measured. These experiments have shown that when, for example, red tomato

extracts are fed labeled lycopene, cyclic carotenoids are produced (Quershi et al.,

1974). This suggests that the enzymes for the production of these compounds are

present, but do not normally accept the lycopene that is produced in vivo. Perhaps

an inhibitor is present in vivo that has been diluted out in vitro. Alternatively, a

spatial separation between these enzymes may exist in vivo that is disrupted in vitro

(Papastephanou et al., 1973). These in vitro studies involved the addition of

quantities of substrate and cofactors that may drastically differ from the situation in

vivo. These unexpected results demonstrate that the pathway for the synthesis of

carotenoids is complicated and a more thorough understanding of the production of

carotenoids requires analysis in less disrupted systems.

Examination of mutants in the synthesis of tomato fruit carotenoids has

predicted that the photosynthetic pigments are produced by different isozymes. A

number of tomato mutants have been studied (for review see Goodwin and Goad,

1970). Many have been mapped, their complement of carotenoids has been measured

and the position of the genetic lesion in the biosynthetic pathway has been predicted

(Goodwin and Goad, 1970; and Khudari, 1972; Tomato Genetic Cooperative Report,

1987). Many of these mutants affect only fruit color, while the photosynthetic

pigments found in the leaves are completely normal (Thelander et al., 1986).

Genetic evidence suggests that fruit carotenoids are produced by two separate

pathways. One pathway produces the abundant amount of lycopene generally

associated with red ripe tomatoes, while the other pathway appears to make a small

amount of B-carotene. Evidence that two pathways exist includes the tomato mutant

apricot, which is deficient in the accumulation of lycopene without affecting the level

of B-carotene in the fruit (Goodwin, 1980). Additional evidence includes the fact that

the production of lycopene is sensitive to high temperatures in normal red tomatoes

(Thelander et al., 1986), while the high f3 mutant has reduced lycopene and enhanced

temperature sensitive synthesis of B-carotene. In other words, the dominant high f.

tomato seems to have used the same temperature sensitive pathway to augment the

levels of B-carotene. Both the existence of the apricot mutant and the residual non-


temperature sensitive amount of B-carotene found in high f fruit imply that a second

pathway exists.

The residual, but ever present, level of total carotenoids found in some

mutants, the presence of temperature sensitive B-carotene, and enzymes that function

in vitro but fail to accumulate their product in vivo suggest that a major ripening

enhanced inducible pathway for carotenoid biosynthesis may exist along with a minor,

possibly constitutive, path. If the minor pathway is responsible for the production of

the cyclic carotenoids, this predicts that enzymes for the production of cyclic

carotenoids may not necessarily be enhanced during ripening in normal red


Characterization of carotenoid biosynthesis in bacteria: a model for higher plant

The cloning and subsequent complementation of genes for carotenoid

biosynthesis in bacteria demonstrates how powerful such a systematic analysis can be

in the understanding of how the pathway functions. These genes are organized into

clusters in the bacteria Erwinia herbicola and Erwinia uredovora. In addition, the

carotenoids in these species are similar to the pigments that accumulate in higher

plants. A number of investigators have used these genes to transform E. coli,

Agrobacterium and Zymomonas (Misawa et al., 1990; Hundle et al., 1991; Misawa et

al., 1991; Naagawa and Misawa, 1991). Deletion analysis of the cluster transformed

into E. coli allowed an unambiguous ordering of the pathway by analysis of the

carotenoids that accumulated in the resulting mutants (Misawa et al., 1990).

The power of the characterization of carotenoid biosynthesis in E. coli is

demonstrated by the analysis of individual genes from different species by Linden et

al. (1991). A plasmid containing the genes responsible for the synthesis of phytoene

in Erwinia uredovora can be transformed into E. coli and sequentially cotransformed

with a plasmid containing a phytoene desaturase (PDS) gene. The type of pigment

that is produced by the specific PDS enzymes can then be monitored. The PDS

cDNAs cloned from Synechococcus, Rhodobacter and Erwinia uredovora were tested.

Analysis of the carotenoids produced in these transgenic E. coli showed that the

different enzymes produced a different complement of products. For example, the

Synechococcus enzyme was able to introduce two double bonds in phytoene while the

Rhodobacter enzyme introduces three double bonds. This analysis also shed some

light on the level of cis-trans isomerizations that accumulate in the various

complemented bacteria. The authors concluded that a gradual shift from cis to trans

carotenoids occurs with an increase in the number of conjugated double bonds

produced by the different enzymes. This suggests that a specific cis-trans isomerase

may not be required for this interconversion. Finally, since the Rhodobacter PDS

adds 3 double bonds, the specificity of the Erwinia lycopene cyclase and beta-

carotene hydroxylase on the asymmetrical molecule formed could be assessed. This

experiment was performed by the introduction of a plasmid containing the entire

Erwinia gene cluster except for a mutant PDS gene. Cotransformation of this plasmid

with theRhodobacter PDS created unique carotenoids. This experiment demonstrated

that the cyclase or the hydroxylase recognizes half of the carotenoid molecule. This


type of analysis can allow a careful identification of the number and types of enzymes

required to produce a specific complement of carotenoids. The sequential cloning

and analysis of the different carotenoid biosynthetic genes in transgenic bacteria

demonstrate how powerful such an approach could be for the dissection of this

pathway in higher plants. If the higher plant carotenoid genes were cloned, a similar

series of experiments could be undertaken.

Molecular biology of carotenoid biosynthesis in higher plants

The gene for phytoene synthesis has been isolated from tomato fruit.

Originally, the clone for the phytoene synthase cDNA (pTOM5) was isolated by

Slater et al. (1985) simply because it recognized increasing amounts of RNA as

ripening progressed. Significant homology to the deduced amino acid sequence of the

tomato cDNA, pTOM5, to a clone for phytoene synthase in Rhodobacter was

identified (Armstrong et al., 1990a). Bartley et al. (1992) used this information to

clone a similar cDNA from tomato fruit and provided evidence, by complimenting

a mutant Rhodobacter, that it indeed coded for phytoene synthase. Recently, a

normal red variety of tomato has been transformed with a constituitively expressed

antisense version of pTOM5 (Bird et al., 1991). This resulted in yellow fruit and pale

flowers. Chlorophyll a and b levels were normal in the leaves of the transformed

plants, which suggested that the leaf carotenoids were unaffected by the constitutively

expressed antisense message. These results confirm that pTOM5 codes for a

phytoene synthase that is specific for ripening fruit. A more complete analysis of the

transgenic plants by Bramley et al. (1992) confirmed that antisense pTOM5 alters

carotenoid accumulation in the flower and fruit but not the leaves of transgenic


The gene for geranylgeranyl pyrophosphate (GGPP) synthase has been cloned

from pepper fruit (Kuntz et al., 1992). The enzyme was purified, antibody was

produced, and a cDNA was isolated from a ripe pepper fruit expression library. Both

the mRNA and the protein increased during fruit ripening. Immunological

examination has shown that this stromal enzyme is mainly found concentrated

around chromoplast plastoglobuli (Cheniclet et al., 1992).

The deduced amino acid sequence from the cyanobacterium Synechococcus

PDS allowed the cloning of a cDNA from soybean cotyledon by Scolnik's lab

(Bartley et al., 1991). Similar work by Pecker et al. (1992) produced a PDS from

ripening tomato fruit. An interesting observation resulting from this work is that,

while the cyanobacterium and higher plant PDSs share more than 65% deduced

amino acid identity, no homology beyond a short amino terminal region that contains

a conserved dinucleotide-binding motif exists with the fungal, Rhodobacter or Erwinia

PDS. This suggests that separate PDSs may have evolved for higher plants and green

algae in contrast with other photosynthetic organisms. Following a different

approach, Camara's lab purified a phytoene desaturase from Capsicum and used

antibody to isolate a cDNA clone (Hugueney et al., 1992). The deduced amino acid

sequence of that clone has striking homology to the tomato fruit sequence. Only

slight increases in the steady-state level of RNA that hybridizes to this clone occur

at MG. This is in contrast with the analysis of the protein and the enzyme activity,

which increases significantly with ripening. It appears from the evidence provided by

both laboratories (Hugueney et al., 1992; Pecker et al., 1992) that the quantity of this

mRNA is low during ripening. Perhaps during chromoplast development large

increases in mRNA occur only for the early enzymes involved in carotenoid


Mayer et al. (1992) have purified an oxidoreductase that is specific for

chromoplasts and crucial for desaturation of carotenoids. The peripheral nature of

the involvement of this enzyme in carotenoid synthesis is supported by the fact that

antibody to this protein does not inhibit the synthesis of carotenoids. However,

oxidized quinones are required for carotenoid desaturation, so it is assumed that this

enzyme acts to affect the redox state of the chromoplast membranes where a large

amount of oxidized quinones and tocopherols found. Perhaps the oxidoreductase and

the increased amount of antioxidants in pepper chromoplasts point to an

environment that contains an escalating amount of reactive oxygen (Romer et al.,

1992). The desaturation of phytoene requires the transfer of electrons to oxygen.

This might cause an increase in reactive oxygen and lead to the induction of enzymes

that generate antioxidants.

Two pathways are predicted for the synthesis of carotenoids in ripening

tomato fruit. None of the genes for the enzymes responsible for steps in the pathway

beyond the synthesis of zeta-carotene have been cloned. It is interesting that only

minor increases in the steady-state level of RNA for PDS occur during fruit ripening.

Since large increases in steady-state levels of RNA accumulate for both phytoene

synthase in tomato and GGPP synthase in pepper fruit (Slater et al., 1985; Kuntz et

al., 1992), perhaps only the very early steps of this pathway are transcriptionally

induced during ripening, while transcript for the later steps will be constitutively

expressed at a low level. Conceivably the dominant mutants del+ and b+, which

produce increased amounts of 8 and B carotenoids respectively, will contain increased

levels of RNA for the the enzymes responsible for these later steps in the

biosynthetic pathway.

Chloroplast Senescence

Chloroplast senescence involves the degradation of chlorophyll and the

dismantling of photosynthetic membranes. Considering that the analysis of

chloroplast breakdown during chromoplast formation has relied on ultrastructural

analysis and characterization of mostly the decrease in steady-state levels of

photosynthetic proteins, perhaps an understanding of how this process occurs during

leaf sensecence will allow predictions for the plastid transition in ripening fruit.

Indeed, 7 out of 12 fruit ripening enhanced clones share homology to transcripts that

increase in senescing leaves (Davies and Grierson, 1989). During ripening and

senescence starch, chlorophyll and photosynthetic proteins degrade (Davies and

Grierson, 1989). A burst in ethylene synthesis has been correlated with senescence

in leaf tissue (Davies and Grierson, 1989) and respiration also increases.

Photosynthetic protein synthesis during leaf senescence

A decrease in the synthesis of photosynthetic proteins occurs during

chloroplast senescence (Roberts et al., 1987). This loss of proteins follows a similar

pattern as described for chromoplast development by Piechulla. The synthesis of

protein in different ages of Phaseolus leaves was measured, and the amounts of

cytochrome f, cytochrome b6, a and b subunits of the ATPase, 68kD PSI protein, and

LHCP decline in leaves. However, the 32 kD protein persists throughout senescence.

Ultrastructural analysis of chloroplast senescence

Plastid senescence is a gradual ordered process, which occurs within the

plastid. It was initially assumed that chloroplasts were engulfed and degraded within

vacuoles (reviewed by Gepstein, 1988). These early studies concluded that senescing

leaves actually contained fewer plastids. However, reappraisal of this work suggested

that the isolation of fewer plastids in senescing leaves was due to the increased

difficulty in isolating these fragile organelles (reviewed by Gepstein, 1988).

Additional ultrastructural analysis showed that plastids in senescing leaves undergo

a transition in which thylakoids become unstacked and plastoglobuli accumulate

(Gepstein, 1988). This suggests that the degradation of plastid components during

senescence occurs within the plastid itself (Thayer et al., 1987), and is gradual and

ordered. The plastid envelope maintains its integrity until very late in senescence,

long after the degradation of thylakoid membranes is complete (Thomson and Platt-

Aloia, 1987). Furthermore, this means that if enhanced degradation of plastid

components occurs during sensecence then it is possible that new proteins would be

imported into the plastid for this purpose.

The degradation of chlorophyll during chloroplast senescence

The degradation of chlorophyll has been suggested to be the trigger for the

proteolysis of the protein components of the thylakoid membranes. Unfortunately,

the pathway for the catabolism of chlorophyll is not well understood. Work with the

non-yellowing mutant of the temperate grass festuca suggested that an initial step in

thylakoid degradation involves dephytlation of the chlorophyll molecule while it

remains in the pigment-protein complex (Thomas et al., 1989). A polar green

chlorophyll degradation product can be isolated from purified light harvesting

complexes (LHCs) of these mutant plastids. This degradation product is also

produced in normal plastids exposed to anoxic conditions. The buildup of this

intermediate occurs in the mutant plant because a downstream step in the

chlorophyll catabolism pathway is missing. Since the product could be isolated from

purified LHCs the authors concluded that the first step in the degradation of

chlorophyll occurs within the protein-pigment complexes of the thylakoid membrane.

Chlorophyllase has been isolated from chloroplasts and could perform this function

(Matile et al., 1989). However, the protein does not seem to increase during

senescence. Alternatively, the enzyme could remain latent in the plastid and be

activated when necessary. Perhaps the activation of the chlorophyllase requires a

newly synthesized protein?

The next step in the catabolism of chlorophyll is probably an oxidation.

However, it is not clear whether this occurs in the plastid or vacuole (Matile et al.,

1989). One theory suggested that a thylakoid associated chlorophyll oxidase is

induced by an increase in linolenic acid during senescence (Luthy et al., 1984), which

triggers the catabolism of chlorophyll and the resulting cascade of degradative

reactions. This idea may require reappraisal in light of the fact that oxidized pyyroles

have been found in vacuoles and not isolated from senescing plastids (Matile et al.,


Degradation of the thylakoid membranes during chloroplast senescence

The analysis of stay-green mutants in soybean and the previously described

mutant in festuca demonstrates that the degradation of the thylakoid membrane is

mediated by at least two separate pathways and is the result of both nuclear and

plastid-encoded proteins.

For example, the cytoplasmic mutant cytG of soybean specifically retains the

(LHCs) (Guiamet et al., 1991). Clearly the breakdown of the core proteins involves

a separate path from the degradation of the LHCs. Since chlorophyll a and b are still

catabolized in this plant, these data suggest that a plastid-encoded protease is lacking

and is involved in LHC breakdown.

When two homeologous recessive nuclear-encoded genes (did2) of soybean are

present in this tetraploid plant, most of the chlorophyll is retained and all of the

thylakoid proteins appear to be present, in contrast with nonmutant leaves (Guiamet


et al., 1991). These data imply that the degradation of the thylakoid membrane in

soybean is also mediated by nuclear-encoded genes.

Evidence from the mutant festuca underscores the complexity of the

degradation of thylakoids. The phenotype of this recessive mutant is a selective

maintenance of components of the photosynthetic membranes. For example, the level

of LHCP is maintained in the mutant plant long after the amount of this protein has

begun to diminish in normal plants (Thomas and Hilditch, 1987). Additionally, the

degradation of the plastid D1 protein, which normally turns over in either the light

or the dark, is inhibited in the dark, but not the light (Thomas and Hilditch, 1987).

The 33Kd protein, however, degrades normally in this mutant (Thomas and Hilditch,

1987). These mutant plants are also able to photobleach if they are subjected to high

light intensities (Thomas and Hilditch, 1987). These data suggest that two separate

protease systems are responsible for the degradation of thylakoid membranes. One

system is light activated and unaffected by this mutation and another functions in the

dark and is inhibited by this genetic lesion.

Degradation of Rubisco

An examination of the degradation of Rubisco suggests how a stromal protein

may be broken down during senescence. The breakdown of Rubisco can be induced

by oxidative stress, in a process that appears to mimic natural senescence (Mehta et

al., 1992). A specific and highly conserved cysteine is oxidized by the addition of

CuSO4 and leads to dimerization of the protein. This is followed by a transient

accumulation of this stromal protein on the membrane, and, finally, degradation. In

vitro studies have also demonstrated that the oxidized form of Rubisco is inactive and

more susceptible to proteolytic digestion (Penarrubia and Moreno, 1990). This

suggests that the turnover of stromal protein is selective and may involve membrane

associated proteases. It is uncertain whether this response to oxidative damage is

analogous to senescence, however, it does demonstrate that an ordered system of

stromal protein degradation exists in the plastid. It remains to be seen if natural

senescence involves a build up of oxidized protein that enters into this degradation

pathway, or if it is developmentally manifested in some other manner.

Changes in other components of the chloroplast during senescence

A decrease in the level of galactolipids has been noted in senescing

chloroplasts (Woolhouse, 1984). However, Woolhouse (1984) pointed out that it is

uncertain whether this change in lipid content is caused by decreased biosynthesis or

increased degradation.

The isolation of proteases involved in plastid protein degradation has been

fairly unsuccessful (Dalling, 1987). The majority of proteases are found in the

vacuole, and their presence masks an examination of changes in plastid proteases

during senescence.

Control of chloroplast senescence

Overall, plastid senescence requires continued protein synthesis, since the

process can be inhibited by cycloheximide (reviewed by Gepstein, 1988). Although

plastid senescence can be triggered by a number of environmental stimuli such as

darkness, oxidative compounds, or nitrogen deprivation, it is still unclear whether this

process is mediated by an increase in de novo synthesis of specific proteins. In the

case of Rubisco, turnover is slow until a nuclear encoded event (assumed since it is

inhibited by cycloheximide or kinetin) triggers the rapid degradation of this abundant

photosynthetic protein (Thayer et al., 1987).

Perhaps the breakdown of chlorophyll acts as a signal for the degradation of

photosynthetic membrane proteins (Nock et al., 1992). This has been suggested by

the results from studies of the mutant festuca where a mutation in chlorophyll

catabolism inhibits the breakdown of the protein components of the thylakoid

membranes. The production of the dephytlated chlorophyll degradation product

requires protein synthesis (Thomas et al., 1989), which implies that the synthesis of

a cytoplasmic protein is required for induction of chloroplast senescence.

Conclusions about chloroplast senescence

To date, ultrastructural analysis has suggested that the senescence of

chloroplasts is a gradual event within the organelle. The degradation of plastid

proteins is a complex process involving multiple proteases with different properties.

Since the breakdown of chloroplasts is inhibited by cycloheximide, this suggests that

senescence requires continued protein synthesis (Gepstein, 1988). Additionally, the

rapid degradation of Rubisco during senescence implies that this upsurge in protein

loss may require a de novo increase in proteins involved in this degradative process

rather than simply continued constitutive expression.

It is difficult to draw generalizations about what may trigger senescence from

the analysis of the few mutants available. The mutant phenotype could simply result


from the accumulation of particular plastid components due to loss of a single

function and may not represent a protein that is developmentally induced during

senescence. It seems safe to assume that the non-yellowingfestuca demonstrates that

although the thylakoid membranes can degrade in response to photooxidative stress,

the more gradual and perhaps dark associated breakdown requires proteases that are

not active until the pigment component has been modified.

It would be extremely helpful if proteins or the genes responsible for the

senescence of chloroplasts could be isolated. An understanding of their regulation

could suggest how senescence is controlled. It is possible that the conversion from

chloroplast to chromoplast may represent an ideal system for the isolation and

identification of cDNA clones for proteins involved in this degradative process since

the chromoplast transition does not take place in a sensecing tissue.

Heat Shock Proteins and Oxidative Stress

The identification of a chromoplast-associated low MW heat shock proteins

(hsp) during this study provoked an investigation into the literature on why this hsp

may have been induced during the chloroplast to chromoplast transition. One

possibility is that the low MW hsp is induced during certain developmental stages.

Historically, the induction of low MW hsps occurs during times when rapid

restructuring of the tissue is underway (Arrigo, 1987). It is also feasible that the hsp

is induced because the tissue is enduring an oxidative stress. Hsp induction by

compounds that create oxidative damage has been described in animal tissues

(Drummond and Steinhardt, 1987; Cajone and Bernelli-Zazzera, 1988; Courgeon et

al., 1988). Both the induction of hsps during development and oxidative stress will

be considered.

Introduction to Heat Shock Proteins

Heat shock proteins were first identified by their massive induction upon

increases of about 100C above ambient temperature (recently reviewed for plants by

Vierling, 1991; and generally by Jaattela and Wissing, 1992). Hsps have been found

in a diverse range of organisms from bacteria to mammals. These proteins are

broadly categorized by their molecular weight in SDS-PAGE. An individual organism

may contain a number of spatially and functionally distinct proteins from each size

class. For example, a number of different hsp 70 proteins have been described, some

of which are present constitutively (hsc).

The function of these proteins has been best characterized for the constitutive

members of the hsp60 and 70 classes. The 60 kD protein plays a role in the assembly

of the multisubunit Rubisco enzyme complex (Roy, 1989). Involvement in the

assembly of multimeric proteins in the mitochondria has also been shown (Cheng et

al., 1989). The 70 kD proteins have been implicated in moving newly synthesized

proteins from ribosomes to the cellular compartment in which they function (Weber,

1992). The role of the hsps during heat shock has been inferred from the function

of their constitutive forms. It is thought that the hsps bind to proteins denatured by

superoptimal temperatures and may facilitate refolding (Weber, 1992). A role for

hsps in providing thermotolerance has been suggested by experiments that create

protection from severe heat shock by the induction of these proteins with a brief or

moderate increase in temperature (reviewed by Lindquist and Craig, 1988).

The induction of hsp messages is rapid and transient (reviewed by Neumann

et al., 1989). Examination of the steady-state level of RNA for a number of different

hsps in soybean revealed that levels decay more quickly at normal temperatures than

during heat stress and the absolute amount of hsps induced is lower in plants that

have been previously exposed to a small heat stress (Kimpel et al., 1990). This is

reminiscent of the thermotolerance response in that the reaction to heat shock is

reduced by heat pretreatment. The transient nature of this response is apparent in

the fact that the hsp message had already begun to decay while the temperature was

still high, and the message was undetectable within 17 hours after the peak

temperature (Kimpel et al., 1990). The induction of heat shock proteins in field

grown plants has also been demonstrated (Burke et al., 1985; Kimpel and Key, 1985).

Kimpel and Key (1985) found that a transient increase in the message for both low

and high MW hsps occurs under fairly normal Georgia summer field conditions with

soybean. This induction is dramatically enhanced in non-irrigated plants. Burke et

al. (1985) also showed similar responses in material from irrigated versus non-

irrigated fields. However, they compared the canopy temperatures and found that the

non-irrigated field was 100C above the temperature in the irrigated field. This is an

important point since the quantity of hsp message produced increases with increasing

temperature (Chen et al., 1990; Kimpel et al., 1990).

Low MW Heat Shock Proteins

In plants the low MW hsps are generally not present at detectable levels in

leaves under nonstress conditions (Vierling, 1991). The low MW hsps represent a

small multigene family that has been subdivided on the basis of amino acid sequence

similarity (Vierling, 1991). Classes of this family have been found in the cytosol,

plastid, and endoplasmic reticulum. In fact, one class of low MW hsp is induced

during meiosis (Bouchard, 1990). Upon induction, the low MW hsps can accumulate

to massive amounts and form large crystalline deposits within the cytoplasm.

Low MW Heat Shock Proteins and Development

The low MW hsps have been found to increase transiently at different times

during development. The best characterized developmental induction of these

proteins is found during Drosophila embryonic development. Arrigo (1987) suggested

that the period of development in which these proteins are most highly induced

represents a time of massive developmental change when the cell may need carefully

controlled mechanisms for overall transformation. Four low MW hsps (ranging in

size from 22 to 27 kD) follow a complex pattern of expression that appears to be

mediated by the molting hormone ecdysterone. Addition of this hormone to

embryonic cell lines induces these proteins (reviewed in Bond and Schlesinger, 1987).

Deletion analysis of their promoters indicated that regions of the DNA responsive

to hormone and heat induction reside in separate portions of the gene. The same hsp

is, therefore, induced developmentally and by heat stress. Since heat causes

destabilization and unfolding of proteins and embryo development suggests rapid

protein biogenesis, the induction of the same group of proteins during both events

implies that they are involved in either stabilizing or degrading incorrectly folded


Evidence for a developmental change in low MW hsps has begun to

accumulate in higher plants. The 14 and 40 kD hsps are present in dry embryos of

wheat and their mRNA has also been isolated from quiescent wheat embryos not

subjected to heat stress (Helm and Abernethy, 1990). These authors propose that

these hsps may act to protect the seed early in development from the effects of

excess heat. Since the germinating wheat embryo has fewer mechanisms to control

ambient temperature in comparison to an older seedling, maybe the production of

hsps allows the organism to tolerate this transiently harsh environment (Helms and

Abernethy, 1990).

The wheat embryo hsps also demonstrate how the genes for hsps may respond

differently to heat stress at different stages of development. These proteins are heat

inducible at early times during imbibition. However, they are not induced after as

little as 12 hours of imbibition (Helm et al., 1989). The wheat seed is initially quite

tolerant to heat stress, such that pretreatment with a low but hsp inducing

temperature does not improve this thermotolerance (Abernethy et al., 1989). By 9

hours of imbibition, however, a pretreatment can improve the survival of the seed

to heat stress (Abernethy et al., 1989). Another interpretation of these data from

wheat, however, could be that rather than protecting the young embryo from heat

stress perhaps the hsps are present in preparation for development and consequently,

additional induction by heat may be unnecessary.

Messages for low MW hsps have been found in dry embryos in other plant

species. The mRNA to a 17.6 kD hsp is present in dry seed and accumulates during

embryo development of sunflower (Almoguera and Jordano, 1992). This cDNA also

hybridizes to mRNA from 3-day-old seedlings subjected to ABA, mannitol or heat

stress. The mRNA for an alfalfa hsp 18.1 has been found to be present in

microcallus suspension and increases transiently during embryo formation at normal

temperatures (Gyorgyey et al., 1991). This cDNA also hybridizes to RNA from callus

subjected to CdCI2, sucrose and heat stress. These data demonstrate that transcripts

for low MW hsps have been found in plant embryos and appear to be

developmentally produced at nonheat stress temperatures. However, it is unclear

from these data whether the transcript induced by stress is encoded by the same

developmentally induced gene or represents another member of a multigene family.

Plastid-Localized Low MW Heat Shock Proteins

A structural analysis of plastid-localized low MW heat shock proteins may

suggest a possible function. Chen and Vierling (1991) have compared the deduced

amino acid sequences of 5 of these plastid proteins, and have reported three

conserved regions. Two of these regions are also present in the low MW hsps found

in the cytoplasm. A third carboxyl terminal consensus region, which is abundant in

methionine, is unique to the plastid-localized low MW proteins. Computer analysis

of the tertiary structure of the protein in this methionine rich region suggested that

an alpha helix with a hydrophobic and hydrophilic face could form (Chen and

Vierling, 1991). This same type of three dimensional structure is predicted for three

such methionine rich regions of the SRP 54 (signal recognition particle), which is

involved in transporting newly synthesized proteins in the endoplasmic reticulum

(Lutke et al., 1992). Careful structure/function analysis of this region in SRP 54 and

its interaction with signal peptides using crosslinking reagents, antibodies produced

against specific fragments of the SRP 54 and deletion analysis of this protein

revealed that binding of signal sequences of nascent proteins occurs here (Lutcke et

al., 1992). Binding experiments with the 7S RNA showed that this same region is

involved in interaction with the RNA. High and Dobberstein (1991) suggested that

the SRP 54 interacts with signal peptide on the methionine rich hydrophobic face

and with RNA on the hydrophilic face. It is tempting to speculate that the low MW

plastid-localized hsp is involved in a similar protein-protein interaction. Perhaps the

need for these hsps is increased during times when excessive incorrectly folded

proteins are present.

The sublocalization of the plastid-localized low MW hsp has been

controversial. Work by Kloppstech and colleagues suggested that this hsp becomes

associated with thylakoid membranes upon heat stress (Kloppstech et al., 1985).

However, this occurs only with plants grown at low light intensities (Glaczinski and

Kloppstech, 1988). Binding of hsp to the membrane occurs most abundantly at a light

intensity below the level at which net carbon is fixed. Considering the transient

nature of these proteins, and that their transcription usually requires heat stress, this

combination of circumstances could occur at dusk in summer or on the floor of a

tropical rain forest. Chen et al. (1990) have shown that under heat stress and higher

light levels this hsp remains predominantly in the stroma. They reported that the only

plastid-localized low MW hsp from pea associates into 200 kD stromal complexes.

Clarke and Critchley (1992) showed that one of the two low MW plastid-localized

hsps, the 32 kD form, from heat stressed barley leaves associates into 250 to 265 kD

complexes within the stroma and contains no RNA, DNA or the a-subunit of the

Rubisco binding protein. This protein is reminiscent of an hsp 28 found in soybean

plastids, which also contains two low MW hsps of 22 and 28 kD (Clark and Critchley,


The idea that the methionine rich region could interact with denatured

proteins during heat stress suggests that these complexes could sequester unfolded

proteins. The hsp 60 also associates into an 800-900 kD oligomer of 14 subunits

(Roy, 1989). Perhaps the role of the plastid-localized low MW hsp will be easier to

resolve if, like the hsp 60 and 70s, a constitutive or developmentally associated form

can be identified. Maybe the chromoplast-associated low MW hsp described within

this work is induced by the degradation of proteins for the photosynthetic apparatus

that occurs between mature green and intermediate stages of fruit ripening.

Induction of Heat Shock Proteins By Other Stresses

Hsps can also be induced by heavy metals, ethanol, amino acid analogues and

glucose starvation (Vierling, 1991). Arsenite and cadmium produced the greatest

induction of hsps (Czarnecka et al., 1984; Edelman et al., 1988). At lower

concentrations of these compounds, the response, like the heat stress response, was

also transient. By 12 hours of exposure, the messages had begun to diminish

(Edelman et al., 1988).

Oxidative stress and the induction of hsps

The initial correlation between heat and oxidative stress was made in an

examination of Drosophila salivary glands. Puffs, representing high levels of

transcription, may occur at the same loci along the chromosome when either a heat

or oxidative stress is applied (reviewed by Loven, 1988). A combination of heat

treatment and application of an oxidizing agent can act synergistically to reduce the

deleterious effect of a subsequent heat stress (Issels et al., 1987). This

thermotolerance developed from oxidative stress inducers can be blocked by the

addition of catalase during the induction period (Issels et al., 1987).

Studies in Drosophila cells (Drummond and Steinhardt, 1987; Courgeon et al.,

1988) and rat hepatocytes (Cajone and Bernelli-Zazzera, 1988) show a mild induction

of some hsps by oxidative stress. In addition, different hsps were induced by different

chemicals (Cajone and Bernelli-Zazzera, 1988). Hsps are also induced by oxidative

stress in Drosophila cells (Courgeon et al., 1988) where hsp 70 and hsp 23 were

induced by hydrogen peroxide. These experiments point to the conclusion that

oxidative stress and heat shock, which induce many of the same proteins, do not

provoke an identical response.

Heat stress and the induction of free radical scavenging enzymes

It still is not entirely clear how heat and oxidative stress are related. Heat

shock uncouples oxidative phosphorylation and this creates an increased level of

reactive oxygen species (Loven, 1988). Additionally, a link has been made between

the application of a heat stress and an induction of free radical scavenging enzymes.

A hyperthermia induced increase in SOD, CAT and glutathione levels occurs with

exercise training of skeletal muscle (reviewed in Barja de Quiroga, 1992). The

implication is that these antioxidants are induced by mild heat stress produced during

exercise. These additional antioxidants reduce the long term deleterious effects of

free radicals produced in this tissue from the uncoupling of oxidative


The heat induced induction of plant SOD was also noted by Tsang et al.

(1991) in tobacco. The cytosolic form of SOD was induced most dramatically

compared with the mitochondrial or plastid form in tobacco. This infers that plants

also may respond to heat stress by inducing enzymes involved in oxidative stress.

Conclusions about the relationship of heat and oxidative stress

Since both heat and oxidative stresses produce a similar response, a number

of hypotheses have been developed to explain how they are related. Some authors

have suggested that the deleterious effects of heat stress are due solely to oxidative

damage (Loven, 1988). This idea may be a bit extreme in view of the fact that hsps

can be induced under anoxia in Drosophila cells (Drummond and Steinhardt, 1987).

Rather than the induction of antioxidants, therefore, this type of assertion requires


that direct oxidative damage be measured, which can be done by examining the

amount of lipid peroxidation. However, few studies have been preformed in which

hyperthermia is directly correlated with oxidative stress (Barja de Quiroga, 1992). A

corollary to this idea was developed in Ame's lab (Lee et al., 1983), in which the

synthesis of bisnucleoside polyphosphates, also termed alarmones, acts as the signal

for oxidative stress induction of hsps. Apparently an increase in these compounds is

observed during both of these stresses. However, it was soon demonstrated that the

synthesis of alarmones is inhibited by heat stress rather than induced. A buildup in

these compounds occurs because their degradation is actually inhibited by heat.

Addition of alarmones to cells was also shown to have no effect on the induction of

hsps and hsps could actually be induced without an increase in the amount of these

compounds (Guedon et al., 1986).

The hypothesis that is currently favored for the induction of hsps by oxidative

stress is that both heat and free radicals create aberrant proteins that lead to the

production of hsps (Neumann et al., 1989). This probably explains why the two stress

responses do not completely overlap. Since hyperthermia creates unstable proteins

and uncouples oxidative phosphorylation, the increase in free radicals and the

resulting response may occur secondarily. Oxidative stress may mainly induce free

radical scavengers, which can forestall the production of large numbers of aberrant



The molecular characterization of fruit ripening has shown that the events that

characterize the maturation of tomato fruit involve the de novo synthesis of a subset

of ripening associated transcripts. At this point, however, only a few genes involved

in this process have been identified and their mode of regulation assessed.

Even though the conversion of chloroplasts into chromoplasts involves the

degradation of chlorophyll and the breakdown of thylakoid membranes along with

a rapid synthesis of carotenoids, only a small number of genes have been identified

that encode proteins that may be involved in this transformation. To date, two genes

in tomato fruit and one in pepper that code for proteins involved in carotenoid

biosynthesis have been isolated. In addition, only a single gene for synthesis of

cysteine synthase has been identified in pepper chromoplasts. Therefore, it is unclear

how many proteins may be involved in this process and whether their synthesis

increases during development.

Historically the characterization of carotenoid biosynthesis in higher plants

involved following the incorportaion of radiolabeled precursors into crude enzyme

preparations or the identification of the compliment of carotenoids produced by

mutants in the biosynthetic pathway. This type of examination was complicated by

the presence of multiple pathways and by the complexity inherent in assays of

disrupted tissue. The analysis of carotenoid biosynthesis, as demonstrated in bacteria,

in which individual genes can be cloned and the activity can be assessed by

complementation in E. coli may now be possible in higher plants. In this way, an

unambiguous assessment of the role an individual enzyme plays on the synthesis of

carotenoids can finally be evaluated. The determination of how many enzymes are

involved in the biosynthesis of fruit carotenoids can be determined. In addition, the

cloning and characterization of the genes involved in this pathway will allow an

understanding of how the synthesis of carotenoids is controlled during chromoplast


The senescence of chloroplasts in leaves involves the breakdown of

photosynthetic structures, in a gradual process within the plastid. It is clear from

evidence accumulated from the analysis of stay-green mutants that the degradation

of thylakoid membranes involves at least two different systems of protein

degradation, portions of which are encoded in both the nucleus and the cytoplasm.

However, it is uncertain whether de novo synthesis of new proteins occurs during

chloroplast senescence in the leaves.

During the transition of chloroplast to chromoplast in pepper chromoplasts,

an increase of compounds that scavenge free radicals occurs. Since the induction of

hsps has been observed during oxidative stress, it is possible that the induction of an

hsp during fruit ripening is caused by such an increasingly harsh environment.

Clearly, the ripening-associated hormone ethylene can induce SOD in tomato and

tobacco leaves. However, the evidence for an increase in enzymes involved in

oxidative stress in tomatoes remains contradictory. More information is needed

before conclusions can be drawn about what role oxidative stress plays during tomato

fruit ripening.

Clearly the transition within the developing chromoplast involves a rapid

degradation of proteins. The period between mature green and yellow fruit appears

to be one in which a particularly rapid degradation of plastid proteins occurs. Since

an increase in hsps has been associated with developmental stages where a great deal

of structural change occurs, the chloroplast to chromoplast transition may be an ideal

environment for the induction of these proteins.



The transformation of chloroplasts into chromoplasts results in a dramatic

alteration of organellar structure. The thylakoid membranes are dismantled in a

ordered process while carotenoids build up within the plastid (Thelander et al.,

1986). The structural changes that occur during this transition have been documented

by electron microscopy (Harris and Spurr, 1969). However, it is unclear how this

event is regulated.

One of the objectives of this project was to determine whether the

development of chromoplasts during tomato fruit ripening is mediated by an increase

in nuclear-encoded transcripts of chromoplast-associated proteins. Since plastid

transcription and translation is greatly reduced during chromoplast development

(Kobayashi, 1991) we have focused on proteins that are nuclear-encoded, translated

in the cytoplasm and imported into the plastid. The import of proteins into plastids

is highly specific and requires an amino-terminal transit peptide, which is necessary

and sufficient for targeting into the organelle (Keegstra, 1989). We have monitored

chloroplast import of in vitro translation products from RNA of tomato fruit isolated

at different ripening stages. We have also examined translation products from

transcripts encoded from cDNAs (slater et al, 1985) whose steady-state RNA level

increases during ripening.

Materials and Methods

Plant Material

Tomato plants (Lycopersicon esculentum Mill. cv. Rutgers) were grown in a

greenhouse in Gainesville, Florida. Fruit were harvested at four ripening stages.

MG1 was characterized by full-size green fruit with firm locular tissue; MG3 fruit

had completely developed locular gel; the 20-50% ripe fruit were red over less than

half of the outer pericarp; and >80% ripe fruit were red over 80% of the outer

pericarp. The pericarp was frozen in liquid nitrogen immediately after harvesting.

Preparation and Translation of RNA

Total RNA was extracted by combining the protocols of McCarty (1986) and

Grierson et al.(1985). Briefly, this involved grinding the frozen tissue with a coffee

grinder and homogenizing in extraction buffer with a polytron. The extraction buffer

consisted of 100 mM Tris-HCl pH 9.0, 200 mM NaCI, 5 mM DTT, 1% SDS and 20

mM EDTA. The insoluble material was pelleted and reextracted with 1/2 volume of

extraction buffer. The supernatant was extracted 2 times with phenol/chloroform

(1:1) and one time with chloroform. The final supernatant was brought to 0.1M KCI

and precipitated overnight with 2 volumes of ethanol at 40C. The precipitate was

washed several times as described in Grierson et al. (1985) and the soluble material

was applied to a cellulose (Sigmacell type 50) column to remove substances that

interfere with extraction of poly (A)' RNA (Grierson et al., 1985). Poly (A)* RNA

was isolated from total RNA by one passage over oligo(dT)-cellulose (BRL).

Individual pTOM clones pTOM 5, 25, 36, 41, 92, 111, and 114 (Slater et al.,

1985) were subcloned into the in vitro transcription vector pSport (BRL) at the PstI

site. Plasmid DNA was isolated by the alkaline lysis method of Lee and Rasheed

(1990). The plasmid DNA was linearized and transcribed with either SP6 polymerase

or T7 polymerase as described in Cline (1988). The poly (A)+ RNA and the in vitro

transcribed RNA were translated (see Figure legends for specific experiments) with

either a wheat germ system (Cline, 1988) or a rabbit reticulocyte lysate system from

Promega (Madison,Wi.) in the presence of [35S]-methionine.

Chloroplast Isolation and Import Assay

Chloroplasts were isolated from pea (Pisum sativum L. cv Laxton's Progress

9) shoots by a combination of differential centrifugation and Percoll gradient

centrifugation (Cline et al., 1989). Import assays were conducted as described by

Cline (1988) except that translation products were diluted 1:3 with unlabelled

methionine in import buffer; 50 Vl of the diluted translation products were added to

100 pil of chloroplasts (1 mg chlorophyll/ml). Import assays were conducted for 20

min at 250 with white light and 10mM Mg-ATP. After incubation, intact chloroplasts

were reisolated by centrifugation through 35% Percoll with or without a treatment


of the protease thermolysin to remove surface bound proteins (Cline, 1988).

Repurified plastids were lysed in 30 pil of 10mM Hepes/KOH pH 8.0. Experiments

testing the requirement of ATP were conducted in foil wrapped tubes to maintain


Subfractionation and Treatment of Plastids after Import

In experiments examining import of translation products from total poly (A)+

RNA, the repurified plastids were lysed in 10mM Hepes/KOH pH 8.0 and

subsequently subfractionated into stromal (soluble) and membrane (pellet) fractions

by centrifugation for 10 min at 10,000 x g. In experiments with translation products

from individual cDNA clones, the soluble fraction was isolated from the membranes

by lysing the plastids in 10mM Hepes/KOH pH 8.0 on ice for 5 min and centrifuging

30 min at 37,500 x g. The membranes were washed one time with 1 ml of 10mM

Hepes/KOH pH 8.0 and resuspended in the same solution. Alternatively, the

membrane fraction was washed with 0.1 N NaOH (Cline, 1986). Envelopes were

separated from thylakoid membranes as described in Cline (1986), however, 1.3 ml

of plastids (1 mg chlorophyll/ml) were used for each import reaction.

Sensitivity of the imported proteins to thermolysin was tested as follows: A

standard import assay was carried out and divided into two equal fractions. These

fractions were either resuspended in import buffer or lysed with 10mM Hepes/KOH

pH 8.0. Aliquots of 25 pl (1mg chlorophyll/ml) were incubated with 0 or 20 pg of

thermolysin for 45 min at 40 C. EDTA was added to 5mM and the samples were

immediately frozen. An equal volume of 2X-SDS PAGE buffer (Cline, 1986) was

added to freshly thawed samples and the tubes were heated to 980 C for 3 min.

Analysis of Import Products

Import products from total poly (A)+ RNA were examined by SDS-PAGE

and fluorography on 7.5%-20% gradient gels (Cline, 1986). Gel lanes received

equivalent amounts of protein as determined by chlorophyll estimation of the

recovered plastids. Import products from in vitro transcribed RNA were examined

on 12.5% SDS PAGE gels.

Results and Discussion

Profiles of in vitro translation products of poly (A)+ RNA differed

markedly between ripe fruit and mature green tissue (Fig.3-1A). The profiles of the

two later ripening stages closely resembled one another, as did the translation

products from the two samples of mature green tissue. Speirs et al. (1984) reported

a similar result from a more detailed analysis of the polypeptide profiles of total

mRNA from ripening tomato fruit.

Pea chloroplasts were chosen to analyze import into plastids because of their

ease of isolation. The isolation of tomato chromoplasts has been reported by others

(Iwatsuki et al., 1984; Bathgate et al., 1985; Hunt et al., 1986; Wrench et al., 1987),

but we were unable to isolate pure and intact fruit chromoplasts that withstand the

handling required to demonstrate successful import. However, in vitro import of an

amyloplast specific protein into chloroplasts (Klosgen et al., 1989) and the in vivo

import of a chloroplast protein into non-green plastids using transgenic plants

(deBoer et al., 1988) implied that chromoplast proteins should also be imported into

pea chloroplasts.

Changes in the profile of imported proteins during chromoplast development

were detected (Fig.3-1B). A number of proteins decreased during this period (Fig.

3-1B). For example, a 30kD and a 14kD soluble fraction protein diminished as

ripening progressed. Two imported translation products dramatically increased in the

later ripening stages. A 26 kD protein fractionated with the soluble plastid proteins,

whereas a 17 kD protein was associated with the membrane fraction. These results

imply that there is synthesis and import of new proteins during the transition from

chloroplast to chromoplast and that the plastid conversion may be an active

developmental program rather than a simple decline in synthesis of the

photosynthetic apparatus.

We next asked whether an import assay could be used to identify individual

clones coding for chromoplast-targeted proteins. A specific cDNA clone, pTOM5,

was selected to evaluate this system. Slater et al. (1985) originally isolated this cDNA

because its RNA becomes more abundant as ripening proceeds. Armstrong et al.

(1990) recognized that the deduced protein sequence shares approximately 25%

identity with the predicted polypeptide of a bacterial prephytoene synthase. Bird et

al. (1991) confirmed the involvement of pTOM5 in carotenoid synthesis by creating

Figure 3-1. Profiles of in vitro translation products and import products are altered
during tomato fruit ripening. A, In vitro translation of total poly (A)+
RNA with four different ripening stages from tomato fruit pericarp.
1 ug of total poly (A)+ RNA per 50 ul reaction was translated in vitro
in a wheat germ system in the presence of [35S]-methionine.
Translation products were examined by SDS-PAGE and fluorography
on 7.5%-20% gradient gels. (MG1) mature green 1, (MG3) mature
green 3, (20-50%) 20-50% red over the outer pericarp, (>80%)
greater than 80% red over the outer pericarp. B, Import into pea
chloroplasts with in vitro translation products of total poly (A)+ RNA
extracted from tomato fruit pericarp at four different ripening stages.
Import assays were conducted and fractionated as described in
Materials and Methods. The (soluble) fraction was isolated from the
(membranes) of the re-purified lysed plastids by centrifugation for 10
min at 10,000 x g. Equivalent amounts of plastid protein (as
determined by chlorophyll estimation) from the +/- protease treated
plastids were separated on 7.5%-20% SDS-PAGE gradient gels and
fluorographed. ["C]-labelled 10 to 70 kD molecular weight standards
(Sigma) appear as arrows on the left side of the figure. The size of the
two import products were derived from a standard curve and are
designated by arrows on the right side of the figure.


-protease +protease
total total

e e C





~ 5 esA

A 17

36 ,
29 ,

20.1 ,

,low 4w

an antisense pTOM5 tomato transformant that produces yellow rather than normal

red fruit. The transgenic fruit are blocked in the synthesis of phytoene from its

immediate precursor, implying that pTOM5 encodes phytoene synthase (Bramley et

al., 1992). We selected this clone because enzymological data suggest that synthesis

of this nonpigmented carotenoid occurs in plastids (Lutke-Brinkhaus et al., 1982;

Mayfield et al., 1986; Dogbo et al., 1987;).

The major translation product of the pTOM5 transcript was 46 kD (Fig. 3-2,

lane 1); this is in agreement with the size of the hybrid-selected translation product

for pTOM5 identified by Slater et al. (1985). Minor lower molecular weight

translation products were also produced from this transcript and may represent

commencement of translation from internal methionines because these products are

not imported into plastids. Incubation with pea chloroplasts resulted in the

appearance of a 41 kD polypeptide that fractionated with intact plastids and was

resistant to protease treatment of the chloroplasts (Fig. 3-2, lane 4). Protease

resistance was not due to an inherent property of the 41 kD protein because it was

degraded by thermolysin if the plastids were lysed prior to treatment (data not

shown). This result is expected of a protein produced in the cytosol, imported into

the plastid, protected from protease by the organellar envelope, and cleaved of its

transit peptide.

Additional criteria must also be demonstrated for the confirmation of import

into plastids. Import is an energy dependent process (Keegstra, 1989); white light or

+ Protease


Figure 3-2.

The 41 kD form of pTOMS is protected from protease digestion by the
plastid envelope and its appearance requires the addition of ATP.
Lane 1) In vitro transcribed RNA was translated with a rabbit
reticulocyte lysate system in the presence of ["S]-methionine. Import
assays and post assay treatments were conducted as described in
Materials and Methods. Lanes 2-5) Import of pTOM5 in vitro
translation products was conducted with (lanes 2,4) or without (lanes
3,5) 10mM Mg-ATP in darkness. Intact chloroplasts were re-isolated
with (lanes 4,5) or without (lanes 2,3) protease post-treatment.
Proteins from lysed plastids were separated on 12.5% SDS-PAGE gels
and flurographed.

m .m


1 2





exogenous ATP must be provided for the successful import of proteins. Fig. 3-2,

lanes 2 and 3 show that processing of the 46 kD protein into the 41 kD peptide

required exogenous ATP. In the absence of ATP, the 46 kD putative precursor

bound to the chloroplasts but was destroyed by the protease (Fig. 3-2, lanes 3 and

5). The import into plastids should also be time dependent; this was found to be the

case for the production of the 41 kD protein (Fig. 3-3). Thus, the successful import

of the pTOM5 translation product verified the utility of import assays for the

identification of genes encoding chromoplast-targeted proteins.

Pea chloroplasts were fractionated into soluble and membrane fractions to

allow analysis of the sublocalization of the 41 kD protein. The protein distributed

between these two compartments (Fig 3-3). Parallel import assays with the thylakoid

localized light-harvesting chlorophyll a/b protein (LHCP) and the stromal small

subunit of Rubisco (SS) were used to assess the quality of our subfractionation.

LHCP was exclusively found in the membrane fraction whereas virtually all of the

SS was recovered in the soluble fraction (data not shown). During preparation of this

manuscript, Bartley et al. (1992) have reported the isolation of a cDNA from a

tomato fruit library by hybridization with a PCR fragment produced with

oligonucleotides derived from the sequence of pTOMS. They found that their clone

differs from pTOM5 in 19 amino acids that reside 5 amino acids away from the

deduced carboxyl terminal end. They also demonstrated that their in vitro translation

product imports via an ATP requiring process and that their import product also

appears in both stromal and membrane fractions of pea chloroplasts.


0 1 2 4 6 8 10 0 1 2 4 6 8 10




. ... .':

Figure 3-3. The production of the 41 kD form of pTOM5 was time dependent
This protein associated with both stromal and membrane fractions.
pTOM5 import assays were conducted as described in Materials and
Methods. One half of a standard import assay was immediately repurified
on a 35% Percoll cushion after 0, 1, 2, 4, 6, 8, or 10 minutes of
incubation. Chloroplasts were protease treated, repurified and
subfractionated as described in the Materials and Methods. Samples were
analyzed by SDS-PAGE and fluorography.


The membranes of pea chloroplasts were further subfractionated into

envelopes and thylakoid membranes (Cline, 1986); upon fractionation of the

membranes after import of pTOM5, virtually all of the mature-sized protein was

found to be associated with the thylakoid fraction (data not shown). A mild NaOH

wash removed the pTOM5 protein from the membrane (data not shown). This

suggested that the protein was only peripherally associated with this fraction (Cline,


Previous attempts by others to localize the site of phytoene synthesis within

the plastid via enzyme assays have resulted in conflicting conclusions. Lutke-

Brinkhaus et al. (1982) concluded that the production of phytoene is associated with

the plastid envelope. However, Mayfield et al. (1986), and independently, Dogbo et

al. (1987) deduced that phytoene is synthesized in the stroma. The unexpected

association of the protein with thylakoids in import assays and its peripheral

association with the membranes demonstrated by the NaOH extraction may have

provided additional evidence that this protein is functionally associated with the

stroma. Alternatively, perhaps the 41 Kd protein has not folded or localized correctly

in pea chloroplasts. Bartley et al. (1991) reported the import of phytoene desaturase

from soybean leaf into pea chloroplasts. Although this protein imported effectively,

Bartley et al. found that it does not also fractionate into the plastid compartment

expected from biochemical analysis. Immunocytolocalization may clarify these

contradicting results. Therefore, it appears that import assays can be useful for the

identification of chromoplast-targeted proteins; however, it is currently unclear what

conclusions can be made from the in vitro organellar sublocalization when using

heterologous systems.

Six additional ripening enhanced pTOM cDNA clones (pTOM 25, 36, 41, 92,

111, 114) of unknown function (Slater et al., 1985) were tested to determine whether

they encode plastid targeted proteins. These particular clones were selected because

a comparison of the published sizes of the cDNA inserts with the size of their

respective mRNA suggested that they might be full length (Slater et al., 1985;

Maunders et al., 1987). In vitro transcription of the cDNAs was carried out in both

directions because the orientation of the clones within the vector was unknown. In

vitro translation products of these transcript pairs failed to import in four of the six

clones. Fig. 3-4, lanes 1 to 7 shows examples of the translation products and lack of

import products produced from these clones. However, since the transit peptide,

which is required for successful import of the protein, is located on the amino

terminal end, and since in vitro translation can begin from any internally situated

methionine, any cDNAs that produced an in vitro translation product but not an

imported product may simply not include the entire coding sequence. Translation

products from pTOM41 and pTOM111 were imported successfully into pea

chloroplasts (Fig. 3-4, lanes 9 and 11). The imported product from both proteins was

smaller than the translation product, which is consistent with processing of the plastid

transit peptide. These proteins were sensitive to protease treatment (see chapter 4),

but resistant after import into plastids (Fig. 3-4). As with pTOMS, import of

pTOM111 and pTOM41 translation products were ATP and time dependent (see

1 23 4 5 6 7 8 9 1011



Figure 3-4.

Translation products from transcripts of additional pTOM clones were
tested for their ability to import into isolated pea chloroplasts. In vitro
translation products from pTOM41 and pTOM111 imported into
isolated pea chloroplasts. In vitro transcribed RNA was translated with
a rabbit reticulocyte lysate system in the presence of [35S]-methionine.
Import assays were conducted as described in Materials and Methods.
Samples were analyzed by SDS-PAGE and fluorography. In vitro
translation with no added RNA (lane 1). pTOM92 transcribed in vitro
with SP6 polymerase, translated in vitro, (lane 2) and imported (lane
3). pTOM92 transcribed in vitro with T7 polymerase, translated in
vitro, (lane 4) and imported (lane 5). pTOM36 transcribed in vitro
with T7 polymerase, translated in vitro, (lane 6) and imported (lane 7).
pTOM41 transcribed in vitro with T7 polymerase, translated in vitro,
(lane 8) and imported (lane 9). pTOM111 transcribed in vitro with T7
polymerase, translated in vitro, (lane 10) and imported (lane 11).

chapter 4). These data support the idea that the cDNAs do indeed code for plastid-

targeted genes. We have begun sequence analysis (see chapter 4) and antisense

transformation experiments of these cDNAs (see Appendix) to investigate their role

in chromoplast function.

We wondered whether our initial examination of import from total poly (A)+

translation products (Fig. 3-1B) might underestimate the number of chromoplast-

targeted proteins. Therefore, we asked whether the pTOM import products could be

recognized among the total poly (A)' RNA import products. The pTOM5 protein

comigrated with a band shown to increase in the import products from the

membrane fraction of late stage fruit (Fig. 3-5, lanes 5 and 6). It is interesting that

the 26kD soluble import product (Fig. 3-1B) migrates similarly to the pTOM111

import product (Fig. 3-5). However, analysis of these two proteins on several gels

suggested that the pTOM111 product was slightly larger. Consequently, it is unclear

if pTOM111 represents a different isozyme or a completely dissimilar protein. Since

the import of the in vitro translation product of pTOM41 associates with the stromal

fraction, this was compared with the soluble fraction of total poly A' import

products. No protein appears to increase in the later stages of ripening fruit which

comigrates with this band (Fig. 3-5, lanes 2 and 3). We concluded, therefore, that not

all of the chromoplast-targeted proteins could be detected in our analysis of total

poly (A)+ RNA import products.

Figure 3-5.

Import products corresponding in size to those for pTOM111 and
pTOM5 were enhanced in late stage tomato fruit. However, no
polypeptide was seen that coincided to the pTOM41 import product.
pTOM41, pTOM5 and pTOM111 were transcribed in vitro with T7
polymerase. RNA was translated with rabbit reticulocyte lysate in the
presence of [35S]-methionine. In vitro translation, import assays and
subfractionation were conducted as described for analysis of individual
translation products in Materials and Methods. Lane 1) Soluble
fraction of the pTOM111 import product. Lane 2) Soluble fraction of
the pTOM41 import product. Lane 3) Soluble fraction of imported in
vitro translation products of total poly (A)+ RNA from >80% ripe
fruit. Lane 4) Soluble fraction from imported, in vitro translation
products of immature green total poly (A)+ RNA. Lane 5) Membrane
fraction of the pTOM5 import product. Lane 6) Membrane fraction
from imported, in vitro translation products of total poly (A)' RNA
from >80% red ripe fruit. Lane 7) Membrane fraction from imported,
in vitro translation products of immature green total poly (A)+ RNA.

1 2 3 4

5 6 7




It is presently unclear how many plastid targeted proteins increase during the

chloroplast to chromoplast transition. The data presented in Fig. 3-4 and 5 clearly

demonstrate that our original examination of chromoplast targeted proteins shown

in Fig. 3-1 underrepresents the number of proteins that increase during this process.

In fact, this is not surprising when considering the limited sensitivity of import of

total poly (A+) translation products. Therefore, we believe that individual testing of

ripening enhanced cDNA clones, with the inherent increase in the quantity of the

individual protein product, is required to clearly identify all the chromoplast targeted

proteins whose RNA increases during this developmental process.

Little detailed analysis of the regulation of ripening enhanced genes has been

undertaken. Lincoln and Fisher (1988) and DellaPenna et al. (1989) concluded that

increases in transcription rate were mainly responsible for the increases in steady-

state RNA levels for the ripening enhanced genes they have analyzed. However, they

also invoked post-transcriptional mechanisms to explain differences between the rate

of transcription and accumulation. Although we cannot exclude post-transcriptional

mechanisms for the increase in steady-state levels of the RNAs we have analyzed,

the simplest interpretation for our data that is consistent with other ripening

enhanced genes is that the processes involved in the development of the chromoplast

are directed at a transcriptional level in the nucleus.

Utilizing an in vitro assay for the import of proteins into plastids we have

identified two proteins from in vitro translation products of total poly (A)+ RNA that

increase during the ripening of tomato fruit. Additionally, we have shown that the

in vitro translation products from three individual ripening enhanced cDNA clones

import into plastids. Therefore, the transition from chloroplast to chromoplast

coincides with an increase in steady-state levels of RNA for proteins destined to this

organelle. This suggests that this process is not a simple decline in the synthesis of

photosynthetic proteins that causes a breakdown of the thylakoid membranes and a

shunting of substrates into the carotenoid pathway, but an active developmental




As tomato fruit ripens, cell walls soften, hexoses are deposited into vacuoles,

and carotenoids accumulate in developing chromoplasts. The transformation of

chloroplasts into chromoplasts is distinguished by the breakdown of the

photosynthetic apparatus and a massive synthesis and deposition of carotenoids.

Little is known about how the catabolism of thylakoid membranes or the synthesis

of carotenoids is regulated. Only two tomato genes for early steps in the carotenoid

biosynthetic pathway have been cloned (Bird et al., 1991; Bartley et al., 1992; Pecker

et al., 1992). Besides proteins involved in the synthesis of carotenoids, only the

plastidial superoxide dismutase, characterized by Livine and Gepstein (1988), has

been shown to increase during tomato fruit ripening. However, Romer et al. (1992)

have also isolated and cloned a cysteine synthase from pepper chromoplasts. These

latter findings suggest, therefore, that the chromoplast may be responding to an

increase in oxidative stress during development.

In order to understand how the transformation from chloroplast to

chromoplast is mediated one objective of the present study was to identify and clone

genes for proteins involved in chromoplast development. We have identified plastid

proteins that may be increasingly synthesized during chromoplast development by

examining cDNAs isolated from a ripening enhanced tomato fruit library, translating

their transcripts in vitro and testing whether the resulting proteins can be imported

into isolated pea chloroplasts. This chapter describes the characterization of two

cDNAs that code for transcripts whose steady-state level increases during

chromoplast development and whose proteins can be imported into plastids. The

characterization of these two cDNAs represents a novel use of import assays to

examine chromoplast development. Since this type of analysis is free of functional

bias, the identification of unexpected components of chromoplast development is


Materials and Methods

Plant Material

Tomato plants (Lycopersicon esculentum Mill cv. Sunbeam) were grown in the

field at the Horticultural Unit of the University of Florida in Gainesville, Florida.

Fruit were harvested at 9:00 am on June 19, 1992 from five different ripening stages.

Mature green 2 (MG2) was characterized by full size green fruit with fully developed

gel in at least one but not all of the locules; mature green 3 (MG3) fruit had

completely developed locular gel; breaker stage fruit were characterized by the

appearance of carotenoids at the blossom end; 20-50% ripe fruit were red over less

than half but more than 20% of the outer pericarp; and >80% ripe fruit were red

over 80% of the outer pericarp. Pericarp from fruit of each stage was frozen in liquid

nitrogen immediately after harvesting at -700C for analysis at a later date.

For the examination of ripening under controlled temperature, MG fruit were

placed in jars with a flow through system of humidified air and allowed to ripen at

220C in the dark. This prevented water stress or a buildup of respiratory CO2. Fruit

from three different ripening stages (MG3, breaker, and 20-50% ripe) were selected

and exposed for 6 hours to 360C in this humidified environment. Following the

incubation period, pericarp was frozen in liquid nitrogen.

Preparation of RNA

Total RNA was extracted by combining the methods of Chomczyinski and

Sacchi (1987) and Cathala et al. (1983). Briefly, this involved grinding 30 g of frozen

tissue in a coffee grinder and homogenizing in 60 ml of extraction buffer with three

30 sec bursts of a Brinckman polytron. The extraction buffer contained 4M guanidine

isothiocyanate (Bethesda Research Laboratories), 0.5% sarkosyl, 25 mM sodium

citrate pH 7.0, and 0.1M 2-mercaptoethanol. Sodium acetate pH 4.0 was added to

0.18M, followed by one volume of phenol and 1/10 volume chloroform. The

homogenate was shaken for 10 sec after each addition and then cooled on ice for 15

min. The phases were separated by centrifugation at 10,000 x g for 20 min, the

aqueous phase filtered through 2 layers of Miracloth (Calbiochem), and nucleic acids

precipitated with an equal volume of isopropanol for one hour at -200C. The

precipitate was collected by centrifugation for 45 min at 10,000 x g. The pellet was

dissolved in extraction buffer and RNA was precipitated with 5 volumes of 4M LiCI

for 15 hours at 40C. Following centrifugation for 90 min at 10,000 x g, the pellet was

washed with 3M LiCI and recovered by centrifugation for 1 hour. The RNA was

dissolved in 0.1% SDS, extracted with 1 volume of phenol/ chloroform (1:1), then

reextracted with 1 volume chloroform, and precipitated with 0.3M sodium acetate

and 2.5 volumes of ethanol. Precipitated RNA was dissolved in sterile water and

stored at -700C.

Template Preparation, In Vitro Transcription and In Vitro Translation

A cDNA for the hsp 21 from heat stressed pea leaves, pHSP 21, was provided

by E. Vierling (Vierling et al., 1988). The pTOM 111 and pTOM 41 cDNA were

provided by D. Grierson (Slater et al., 1985). The pTOM cDNAs were digested with

Pstl and ligated into pSport 1 (Bethesda Research Laboratories) at the PstI site, and

transformed into E. coli strain TB1. Plasmid DNA was isolated by the alkaline lysis

method of Lee and Rasheed (1990). The pHSP 21 plasmid was linearized with Barn

HI and in vitro transcribed with SP6 polymerase as described in Cline (1988). The

pTOM 41 and pTOM 111 plasmid DNAs were linearized with Eco RI and in vitro

transcribed with T7 polymerase (Bethesda Research Laboratories). The in vitro

transcribed RNA transcripts were translated with a rabbit reticulocyte lysate system

from Promega (Madison, Wi.) in the presence of either [3S]-Met or [3H]-Leu

according to procedures described in the Promega manual.

Northern Hybridization

Total RNA, quantified by spectrophotometric absorbance at A2, was loaded

(15 Ig/lane) onto a formaldehyde denaturing gel (Sambrook et al., 1989). After

electrophoresis the gel was washed in sterile water and stained in ethidium bromide

(l1g/ml) for 30 min and destined in 20X SSC, which contained 0.3 M sodiun citrate

and 3 M sodium chloride, for 45 min. The amount of ribosomal RNA was visually

assessed to ensure that approximately equivalent amounts of RNA were loaded. The

RNA was capillary blotted onto nylon membrane (Amersham Corp.) in 20X SSC

overnight. The blot was washed with 10X SSC and exposed to 1.5 J/cm2 UV light

with a transilluminator (Bios).

For hybridization probes, the entire cDNA insert was excised with PstI,

isolated by gel electrophoresis, and electro-eluted. 100 ng of DNA was radiolabeled

with 75 pCi of a32P-dCTP by random primer labelling with a kit (Boehringer


Prehybridization and hybridization conditions were as described in Church and

Gilbert (1984). The prehybridization buffer contained 0.5M sodium phosphate pH

7.2, 7% SDS and 1% bovine serum albumin (Sigma, fraction V) at 650C. Blots were

prehybridized in 40 ml of buffer for at least 1 hour at 650C. The labeled probe was

denatured at 1000C for 5 min and added to the blot with 5 ml of prehybridization

solution. This was incubated overnight at 65*C. The blot was washed one time in 1X

SSC and 0.1% SDS at room temperature and then 3 times in 0.2X SSC and 0.1%

SDS for 30 min. X-ray film (Kodak, X-OMAT) was exposed to the blot overnight

with two intensifying screens (DuPont) to locate bands. Bands were excised from the

membrane and quantified by liquid scintillation counting.

DNA Sequencing

The PstI excised pTOM 111 cDNA insert was subcloned into pBluescript (SK)

plasmid (Stratagene) at the PstI site. This was restricted with either XbaI and SstI or

KpnI and Eco RV for production of ExoIII/Mungbean nested deletion plasmids

following the protocol described in the manual (Stratagene). The resulting plasmids

were precipitated, resuspended in TE, ligated directly, and transformed into E. coli

strain TB1. Plasmids containing nested deletions from either direction were selected

and sequenced using the T3 or M13-40 primer of the pBluescript vector. Double

stranded plasmid DNA was used as a template and prepared as described by Lee

and Rasheed (1990). DNA sequencing procedures were as described in the

Sequenase sequencing kit manual (US Biochemical).

The pTOM 41 cDNA was blunt end ligated (Sambrook et al., 1989) into

pGEM-4Z (Promega) at the Smal site. Both the pGem constructs described above

and the pSport constructs described in the section on in vitro transcription were used

as template for double stranded DNA sequencing as described above. Initially the

T7 or M13-40 sites of the pSport or pGem vectors were used as primer sites.

Oligomers (15mers) specific to pTOM 41 were synthesized by the UF-ICBR DNA

Synthesis Core and used in subsequent sequencing reactions. Some portions of the

sequence were generated by the ICBR DNA Sequencing Core. Sequences were

analyzed with the University of Wisconsin Genetics Computer Group computer

programs (Devereux et al., 1984). Protein sequence comparisons were performed at

the National Center for Biotechnology Information using the GENINFO(R) program

of the Experimental Blast Network Service (Altschul et al., 1990) made available by

the ICBR Biological Computing Facility. The petunia hsp 21 sequence (Chen and

Vierling, 1991) is available as EMBL accession #X54103. The pTOM66 sequence

(Fray et al.,1990) is available as EMBL accession #X56138. The sequence for the

hypothetical 30.9 Kd protein from E. coli (Post et al., 1992) is available in GenBank

accession #M77236.

Chloroplast Isolation and Import Assays

Chloroplasts were isolated from pea (Pisum sativum L. cv Laxton's Progress

9) shoots by a combination of differential centrifugation and Percoll gradient

centrifugation (Cline et al., 1989). Import assays were conducted as described by

Cline (1988) except that translation products were diluted 1:3 with unlabelled

methionine or leucine in import buffer; and that 50 l of diluted translation products

were added to 100 pl of chloroplasts (1 mg chlorophyll/ml) for a 300 pl total assay

volume. Import assays were conducted for 15 minutes at 250C with white light and

10mM Mg-ATP unless indicated otherwise. Experiments testing the requirement of

ATP for import of pTOM 111 protein involved incubation of 200 pl of the

translation mixture with or without 4 units of apyrase (Sigma) for 5 min at 25C. The

import assays were conducted in foil wrapped tubes to maintain darkness at 250C.

After incubation, intact chloroplasts were reisolated by centrifugation through 35%

Percoll with or without a treatment of the protease thermolysin (Sigma) to remove

surface bound proteins (Cline, 1988). Plastids were treated with 50 tig thermolysin

for 40 min at 4C unless indicated otherwise. Import products were diluted 1:1 with

2X-SDS PAGE buffer, incubated 8 min at 600C and analyzed by SDS-PAGE and

fluorography (Cline, 1986) on 12.5% gels.

In order to demonstrate that the imported protein was sensitive to

thermolysin, an import assay was performed and the chloroplasts were lysed before

the addition of the protease. One fraction of a standard assay was resuspended in

100 jIl import buffer, the other fraction was lysed with the same amount of 10mM

Hepes/KOH pH 8.0. Thirty pl aliquots (0.5mg chlorophyll/ml) were incubated with

0, 12, and 24 pg of thermolysin for 45 min at 40C. For the pTOM 41 protein, [3H]-

leucine was used in the in vitro translation and 1/2 of a standard import reaction (50

pl of plastids at 1mg chlorophyll/ml) was resuspended in 25 pl 10mM Hepes/KOH

pH 8.0. This sample was divided in half and 5 ig of thermolysin was added to an

aliquot. After 40 min incubation at 40C, EDTA was added to a final concentration

of 5mM and the samples were immediately frozen. An equal volume of 2X-SDS

PAGE buffer (Cline, 1986) was added to freshly thawed samples, the tubes were

heated at 980C for 3 min and 10 pc was loaded onto 12.5% SDS-PAGE gels.

For analysis of the location of imported proteins, chloroplasts recovered from

assays were lysed in 50 p1 10 mM Hepes/KOH pH 8.0, and incubated for 5 min.

Stroma was separated from membranes by centrifugation for 30 min at 37,500 x g.

The membranes were washed with 1 ml of import buffer, centrifuged for 30 min at

37,500 x g. and finally resuspended in 40 tl of 10mM Hepes/KOH pH 8.0.


The tomato fruit cDNAs pTOM 111 and 41 were selected by Slater et al.

(1985) because transcripts homologous to these clones increased during ripening.

Data presented in the previous chapter have shown that in vitro transcripts from

these clones can be translated in vitro and the translation products can be imported

by pea chloroplasts. Further studies were performed in order to ensure that all the

criteria for successful import were met, such as the energy and time dependence of

the reaction along with sensitivity of the mature form of the protein to protease.

Import Studies of pTOM 111 Protein

The pTOM 111 in vitro transcript was translated into a 30 kD protein and

processed by the isolated plastids into a 24 kD form in an energy (Fig.4-1A) and

time (Fig.4-1C) dependent reaction. The production of this processed form was

virtually prevented by the addition of apyrase, an ATPase, to the translation mix in

the no ATP assay (Fig. 4-1A lanes 1 and 2 versus lanes 3 and 4). Only the 30 kD

precursor, which remains outside of the plastid, was removed by the protease

treatment (Fig. 4-1A lanes 1 and 4). The mature protein was sensitive to the protease

only after the plastids had been lysed prior to addition of this enzyme (Fig. 4-1B).

+ +
- +
3 4

24 N in

Protease pg
Lane #

jw 4

Min 0 1.5 2 4 6 8 10
Lane # 1 2 3 4 5 6 7


The 30 kD pTOM 111 translation product can be imported into pea
chloroplasts and processed into a 24kD form in an energy and time
dependent process. In vitro transcribed RNA was translated with a
rabbit reticulocyte lysate system in the presence of [3S]-Met. Import
assays were conducted as described in the Materials and Methods. A)
Production of the mature form of the protein requires ATP and is
protected from protease digestion. A standard import assay was
conducted with (lanes 3 and 4) or without (lanes 1 and 2) 10mM Mg-
ATP. Plastids were pelleted and resuspended with (lanes 2 and 4) or
without (lanes 1 and 3) protease. B) The plastid membrane protects
the mature form of pTOM 111 from the added protease. Plastids were
(lanes 4, 5 and 6) and were not lysed (lanes 1, 2, and 3) after import
with 0 (lanes 1 and 4) 12 pg (lanes 2 and 5) or 24 pg (lanes 3 and 6)
of protease. C) The production of the mature form of pTOM 111 is
time dependent. Assays were of 100 pl of plastids (1mg/ml chlorophyll)
for each time point. The reactions were terminated with 3.3 mM HgCI2
to quickly stop the reaction.


12 24
+ +
5 6

Figure 4-1

When HgClz was used to terminate the assays, a small amount of protease-resistant

precursor was observed in the early time points (Fig. 4-1B). This may reflect a stage

in the reaction prior to cleavage of the transit peptide but after import into the


As will be discussed in the following section, sequence analysis identified

pTOM 111 as encoding a plastid-localized low MW heat shock protein (hsp).

Consequently, we asked whether our hsp would localize in vitro in the same manner

as a plastid localized low MW hsp that was isolated from heat stressed pea leaves.

Chloroplasts recovered from import assays were separated into membrane and

soluble fractions in Fig. 4-2. The proteins from both pTOM 111 and the pea cDNA

were recovered in the soluble fraction. Even though the tomato cDNA was isolated

from ripening fruit, the protein behaves in a manner similar to one isolated from

heat stressed leaves.

Import Studies of pTOM 41 Protein

A similar analysis of the import of an in vitro translation product from pTOM

41 was performed (Fig 4-3). In this case, a 47 kD protein was produced upon in vitro

translation of the pTOM 41 transcript (Fig. 4-3B). When the translation product was

incubated with isolated pea chloroplasts, a 41 kD protein was generated (Fig. 3B).

This predicts that a 6 kD transit peptide is cleaved upon import into the plastid.

Little of the 47 kD precursor was associated with chloroplasts recovered from assays

without added ATP, even without protease post-treatment (Fig. 4-3A, lane 1). Since

pTOM 111
Tr St Mem
1 2 3

Figure 4-2.

pHSP 21
Tr St Mem
4 5 6

The pTOM 111 and an hsp 21 from heat stressed peas localize
predominantly to the stroma. In vitro transcript was translated with a
rabbit reticulocyte lysate system in the presence of [5S]-Met and
import assays were conducted as described in the Materials and
Methods. Translation products (Tr), stroma (St) and membrane
(Mem) fractions were analyzed by SDS-PAGE gels and fluorography.

Lane #

- + -
- +
1 2 3

T St Mem

+ + + +
+ + + +
4 5 6 7

- ~ ~

Lane #

01 2
1 23 4

8 10
7 8

41 .m.

Figure 4-3. The 47 kD pTOM 41 translation product can be imported into pea
chloroplasts and processed into a 41 kD form in an energy and time
dependent process. A) Production of the mature form of the protein
requires ATP and is protected from protease digestion. The mature
form of PTOM 41 can be digested by the protease if the plastids are
lysed. Import assays contained 10mM Mg-ATP (lanes 2, 4, 5, 6, 7, 8,
and 9). Thermolysin was added after import (lanes 3, 4, 5, 6, 7, and 9).
Plastids were lysed after import in 10mM Hepes/ KOH pH 8.0 and
treated with thermolysin (lane 9). Stromal (St) and membrane (mem)
fractions were separated from total (T) plastids following an import
reaction as described in the Materials and Methods. B) The
production of the mature form of pTOM 41 is time dependent. Assays
contained 50 gl of plastids (1mg/ml chlorophyll) for each time point.
[H]-leucine replaced [35]-methionine in the in vitro translation

the plastids were repurified before the assays were analyzed, this suggests that little

of the pTOM 41 translation product attaches to the outside of the plastid.

Subfractionation of recovered chloroplasts showed the pTOM 41 imported protein

to be in the soluble fraction (Fig. 4-3A, lane 7). If plastids were lysed after import

and treated with protease, the pTOM 41 mature protein was digested (Fig.4-3A, lane

9 versus lane 8), which demonstrates that it is not inherently resistant to this

treatment A small amount of protease-resistant precursor was also found in the

early time points after import of the pTOM 41 protein (Fig. 4-3B).

These data demonstrate that the in vitro translation products of both pTOM

41 and 111 are imported into the plastid, are protected from protease by the

organellar envelope and are cleaved of their transit peptides. In order to further

characterize these chromoplast-associated proteins, the cDNAs were sequenced and

the corresponding amino acid sequences deduced (Materials and Methods).

The pTOM 111 cDNA Codes for a Low MW HSP

The pTOM 111 sequence (Fig. 4-4) shows significant homology to those of

low MW plastid-localized heat shock proteins isolated from heat stressed leaves.

Proteins from soybean, pea (Vierling et al., 1988) Arabidopsis, petunia (Chen and

Vierling, 1991) and maize (Mieto-Sotelo et al., 1990) have been cloned and

sequenced. The pTOM 111 sequence is 81% homologous to a cDNA isolated from

heat stressed petunia leaves. Low MW hsps from plants have been categorized on


-30 -10 10 30
N A Y T S L T S S P L V S N V 15

50 70 90 110
----------- ----------------......... --------------------------....... ................ .... ---------
15 S V G G T S K I N N N K V S A P C S V F V P S N R R 41

130 150 170 190
42 P T T R L V A R A T 6 D N K 0 T S V D V H H S S A 0 G 68

210 230 250 270
---------- ------.....................................---.......-------.........---------
69 G N N Q G T A V E R R P T R N A L 0 V S P F G V L o P 95

290 310 330 350
........ -----------..------------------------+.........-----------.......... -----------......... ........---------
95 S P M R T M R Q M I D T M D R L F E D T M I P G R 121

370 390 410 430
+----------.... --------... -------.. ------... ---------.. ................--........
122 N R A S G T G E I R T P U D I H D D E N E I K N R F D 148

450 470 490 510
4.---------...-... ........ 4 .-- ..................-----------...---- ......------------------
149 N P G L S K E D V K V S V E N D N L V I K G E H K K E 175

530 550 570 590
.-------- .-------.-----------.-----------4.-------.----------------------.-----------
175 E D G R D K H S W G R N Y S S Y D T R L S L P D N V 201

610 630 650 670
202 V K D K I K A E L K N G V L F I S I P K T E V E K K V 228

690 710 730 750
.---. ---.------ ------.. .---..-----------..... -----------....--........---------
229 I D V Q I N 255

770 790
---------... --------------------..... 4------- ......... -

Figure 4-4. Nucleotide and deduced amino acid sequence of pTOM 111. The
sequence was analyzed using the University of Wisconsin Genetics
Computer Group programs. The position of the specific nucleotides
appears above the sequence while the position of the specific amino
acid appears to the side.

the basis of deduced amino acid sequence similarity (Vierling, 1991). There are three

classes of cytoplasmically-localized low MW hsps, one class associated with the

chloroplast, and another affiliated with the endoplasmic reticulum. A comparison of

the deduced amino acid sequence of pTOM 111 to these proteins revealed the

greatest similarity with the petunia plastid hsp (77% identity) (Fig. 4-5). This is not

surprising since these species are evolutionarily most alike. Comparison to the only

other characterized low MW hsp from tomato fruit, pTOM 66 (Fray et al., 1990), a

cytoplasmically-localized hsp, revealed only 32% identity to pTOM 111 (Fig. 4-5).

The pTOM 66 protein shares more sequence similarity to cytosolically localized hsps

from other species (data not shown). Overall these hsp classes share more sequence

identity among members from different species than they do with hsps from other

families within the same species. The homology of pTOM 111 to a cDNA for a

known plastid-localized protein supports the use of import assays to identify cDNAs

for chromoplast-associated proteins.

Chen and Vierling (1991) have identified 3 conserved regions in 5 plastid

localized low MW hsps they have examined. While the first two regions are found

in all low MW hsps, the third region is unique to the plastid-localized proteins. It is

in these three regions that pTOM 111 has the highest homology to other low MW

hsps. A 90% amino acid identity to the petunia protein is found in these regions in

contrast with the 77% overall homology (Fig. 4-5).



++ ++ + +++++ + ++ ++ ++ ++ + + +
macktttcsa sptvsngws atsrtnnkkt ttapfsvcfp yskcsvrkpa
.......... .......... .......... .......... ..........

Putative Transit Peptide
51 100
+ + .+ + + +
+4-4+ +4"+4 +4+-H- + + +++++++++++ + ++ +4+++
srlvaqatgd nkdtsvdvhv snnnqggnnq gsaverrprr maldvspfgL
NSLIPRIFGD RRSSSN.... .....................FDPFSIDV




+ + +

44+ 4-+H4a +4. 44++.+
Ldpmspnirtm rqmmdtmdrl
FDP....... ..........

+ +

4++++ + -+++- 44+4- 4+4+
fedtmtfpgs rnrgtgeira pwdikddene

sensus Region III

+ +++ + + ++++ +

ikmrfdmpgt skeevkvsve

+ ++ ++44 + +
+ ++++ ++ +++ + + ++++++
dd.vtvikge hk..keesgk d.dswgrys

Consensus Region II

+ +44 + + +
++4+++4++444 4+4 +4 +4- +
sydtrlstpd nvdkdkvkae

4*4+** *+ + +
+++++ +++ 4+4 +4+4 ++ ++ +
Lkngvllisi pktkvek.kv tdveik.

Consensus Region I

Figure 4-5.

The pTOM 111 shares 77% amino acid sequence identity to an hsp 21
from petunia and 32% identity to pTOM 66 from tomato fruit. The
deduced amino acid sequences were compared with the University of
Wisconsin GCG program, "pileup", as described in the Materials and
Methods. The + symbol represents exact matches between proteins,
while dots represent spaces built into the protein by the pileup
alignment The putative transit peptide (Vierling, 1991) is delineated
by a solid underline. The three consensus regions described by Chen
and Vierling (1991) are also marked with a solid line.

The putative transit peptide of pTOM 111 protein is identified in Fig. 4-5.

(Vierling, 1991). Although this part of the protein is the least similar to the petunia

hsp shown in Fig. 4-5 (63% amino acid identity), the region still contains a number

of features characteristic of transit peptides (Keegstra et al., 1989). For example, it

is rich in serine, valine and alanine and deficient in the acidic amino acids glutamic

and aspartic acid. In addition, some basic amino acids (arginine and lysine) are also

present. Where the sequence varies from the petunia hsp within this region the

substitutions are generally those residues characteristic of transit peptides. The

observation that this portion of the protein is least similar to other plastid localized

low MW hsps is not surprising, since an examination of transit peptides for the 5

plastid-localized low MW hsps analyzed by Chen and Vierling (1991) has also shown

that they are the most diverged portion of the protein.

Even though the structure and function of the chromoplast differs drastically

from the chloroplast, a similar hsp is produced. Our data agree with the conclusion

of Vierling (1991) that localization to the plastid appears to predict more sequence

conservation than does the tissue or organism in which the protein is expressed.

Although the increases in low MW hsps pTOM 66 and pTOM 111 may be a general

result of ripening, sequence conservation suggests that their role may be more

pertinent to the compartment in which they function.

Sequence Analysis of pTOM 41

The sequence obtained for pTOM 41 is shown in Fig 4-6. The cDNA is 1423

10 30 50 70
.-..........-- ..+----.............. -....... .......-+...........-
L 1
90 110 130 150
2 W L P V I F F V V 8 N P K L I L L K R V V F F 0 8 W 8 28

170 190 210 230
.........----------------...........................4........ ......... .........4-.........4
29 N R P H G 8 Y F N K N I 0 F R R N 8 F V I V K A 8 G 55

250 270 290 310
.--.. ---...... ...-....- -.............. --...-....-4.......-...- ---.----.........
56 R T K K Q V E I T Y N P E E K F N K L A E V 82

330 350 370 390
-------... .+......-..-.......+.........4.........-........4-.........--.........
82 R E A G L 8 R L T L F S P C K I N V F L R I T 8 K R 0 108

410 430 450 470
.........-.........-.........4.........4........................... 4. ........4
109 D G Y H D L A 8 L F H V I 8 L 0 D K I K F 8 L 8 P 8 K 135
490 510 530 550
.........----------------................... ------.................----.......-... .........4
136 SKD R L T N V A V P L D E R N LII K A L N L 162

570 590 610 630
----......-.........------.... .....-...........4.................------.........
162 Y R K K T 0 T 0 N Y F W I H L 0 K K V P T G A G L G G 188

650 670 690 710
-------------------------*--------- +--------- ---------------
................................................4. ...4..................4
189 G 8 8 N A A T T L W A A N 0 F S G C V A T E K E L Q E 215

730 750 770 790
216 W 8 E I G S D I P F F F 8 H G A A Y C T G R G E V 242
810 830 850 870
----------------------------------* ---------* -----.----*------
242 V Q D I P 8 P I P F 0 I P M V L I K P 0Q A C 8 T A E 268

890 910 930 950
................ ....4.........4.........4.........4....................4
269 V Y K R F Q L D L 8 8 K V D P L 8 L L E K I 8 T 8 G I 295

970 990 1010 1030
---------*---------+--------------------- --------------------
.........-.........4.........4.-......4.4. + 4. 4..........
296 8 Q 0 V C V N 0 L E P P A F E V L P 8 L K R L K Q R 322

1050 1070 1090 1110
......---.--.......---..-------.........-........4......... --..................
322 V I A A G R G Q Y D A V F M 8 G 8 G 8 T I V 6 V 6 8 P 348
1130 1150 1170 1190
.........---------+ ---.........----------......... -----.........4.........4.........-------4..........
349 D P P O F V Y 0 0 E E Y K D V F L 8 E A 8 F I T R P A 375

1210 1230 1250 1270
---....-----.......-..-- ......+.-...... ........4-.........4.-........4-.........
376 N E W Y V E P V G T I D P E F 8 T 8 F D 402

1290 1310 1330 1350
-...-------........- .........-- ........4.-................- -.......4-........-4

1370 1390 1410
---------... .. ..... ............---- -...........------- .. ..----

Figure 4-6. Nucleotide sequence and deduced amino acid sequence of pTOM 41.

Numbers appearing above the sequence refer to the nucleotide

sequence. Numbers appearing to the side of the sequence refer to the

deduced amino acid sequence.

9B 0

_-_ .- o

S o


-o "
-- I 1

_- _- M ,,.
- -o MSs

0 to1 1" 0

-- ,^

4) 4) 0

n- =
-- o

= 1 g1


bases long. Only 1 long open reading frame (ORF) is present, but it starts with a

leucine, not a methionine (Fig. 4-6 and 4-7). This is hard to understand, given that

the pTOM 41 transcript directed production of a 47 kD translation product. A

methionine does occur within 12 nucleotides of the beginning of the long reading

frame; however, two stop codons are encountered within the next 57 bases. Another

small ORF of 28 residues is found two nucleotides 3' of the large ORF. This begins

with a methionine in a favorable initiation context (Kozak, 1991). Two possible

explanations for the observed anomaly are: 1) a sequencing error, or 2) a +1

frameshift during translation. The first possibility was addressed by additional

sequence analysis of the areas in the vicinity of the 5' and 3' ends of the long ORF.

The cDNA sequence in Fig. 4-6 shows that the long ORF starts at 78 bases

and ends at 1280. It is clear that an ORF of approximately this length is required to

produce the protein observed on SDS-PAGE, all other potential reading frames are

littered with stop codons, therefore, if a simple sequence error was generated, it

would be found within the first few hundred bases. The first 290 bases were

reexamined using a number of different modifications to the sequencing reactions.

Figure 4-8 is a compilation of these individual sequencing results. For example, both

strands of the sequence (R1 and Fl) were examined by comparing incorporation of

dITP instead of dGTP to the standard Sequenase double stranded sequencing

reactions. This region of the sequence was also analyzed by the ICBR DNA

sequencing core. They used either Klenow (R4) or TAQ polymerases (F2 and R2)

in generating these sequences. Since the template that was used for most of the