Download PDF

TTEC June 2000
4/5/00
11:15 am
Page 257
REVIEWS
Engineering chloroplasts: an alternative site
for foreign genes, proteins, reactions and
products
Lawrence Bogorad
Plant genetic engineering via the nucleus is a mature technology that has been used very productively for research and commercial biotechnology. By contrast, the ability to introduce foreign genes at specific locations on a chloroplast’s chromosome
has been aquired relatively recently. Certain limitations of nuclear genome transformation methods might be overcome by
the site-specific introduction of genes into plastid chromosomes. In addition, plastids, mitochondria and other subcellular
organelles might provide more favorable environments than the nuclear–cytoplasmic compartment for certain biochemical
reactions and for accumulating large amounts of some gene and enzyme products.
n crown-gall disease, infection by an Agrobacterium
tumefaciens strain that harbors a large tumor-inducing
(Ti) plasmid1,2 causes a mass of mainly undifferentiated plant cells to form on a plant’s stem at the soil
line (the crown). The transfer DNA (T-DNA) portion
of the Ti plasmid and its delimiting ‘right’ and ‘left’
border sequences become integrated into the nuclear
genome of a susceptible plant cell that is in contact with
the bacterium. The T-DNA encodes, among other
things, enzymes for synthesizing plant hormones that
stimulate cell division and the proliferation of undifferentiated cells into the tumor. However, T-DNA
from which the genes for hormone-synthesizing
enzymes have been removed is an excellent vector for
introducing foreign DNA into a nuclear chromosome
of a plant cell.
I
Nuclear transformation
There are several steps involved in A. tumefaciensmediated insertion of DNA into a plant: (1) a foreign
gene is inserted between the borders of the T-DNA
cloned in a small plasmid; (2) this construct is introduced into an A. tumefaciens strain that lacks T-DNA
and the engineered T-DNA inserts into the Ti plasmid;
(3) the A. tumefaciens bacteria carrying the foreign
DNA on the now-disarmed Ti plasmid are put into
contact with plant cells; (4) the T-DNA is transferred
into a plant cell; and (5) the new gene, as a component
of the T-DNA, is incorporated into a chromosome in
the nucleus of the plant cell.
The next step after producing a transformed cell is
to obtain a fertile transgenic plant. Research into plant
tissue culture started at the beginning of the 20th
century. One aspect of this work culminated in the
demonstration that single carrot cells, cultured on an
appropriate medium, can divide, give rise to embryos
and grow into entire plants3. The culture media used
in the initial experiments included complex and poorly
L. Bogorad ([email protected]) is at the Department of
Molecular and Cellular Biology, Harvard University, 16 Divinity Ave,
Cambridge, MA 02138, USA.
TIBTECH JUNE 2000 (Vol. 18)
characterized components but later, using information
acquired in studies of plant-hormone action, entire
plants could be regenerated by transferring cells in culture through a series of culture media with different
hormone contents. Even mature cells from, for example, tobacco leaves could be stimulated to multiply to
form clumps of cells from which fertile plants formed.
The results of these two lines of research – investigations of the Ti plasmid and the regeneration of whole
plants from cells in culture – were used together in the
early 1980s to produce fertile transgenic plants. The
most significant discovery for the genetic engineering
of plants was that if a gene was introduced via T-DNA
into a cell that was then used to generate a fertile plant,
that gene would propagate to tobacco progeny through
the sexual cycle and seed. The commercial potential
for crops was, of course, apparent immediately. At that
time, the major limitation to producing a genetically
transformed crop plant appeared to be the ability to
regenerate that plant from cells or clumps of tissue in
culture. Only a few species could be manipulated in
this way then; now many can be.
Crop plants
A. tumefaciens-mediated nuclear transformation technology has been used widely to create transgenic
dicotyledonous plants, but monocotyledonous plants
proved difficult to transform by this method. The results
of early experiments led to the view that most monocotyledonous plants – including cereals – were beyond
the host range of A. tumefaciens. However, experiments
in the late 1980s showed that maize-streak-virus DNA
can be delivered to maize via the A. tumefaciens T-DNA
system and can be expressed in the plant. It was only
in 1993 and 1994, after the processes involved in TDNA transfer were better understood, that transgenic
rice was produced via A. tumefaciens4. The efficient
transformation of maize via A. tumefaciens was described
in 1996 (Ref. 5) and of wheat in 1997 (Ref. 6). Meanwhile, however, and largely driven by the prospect of
crop improvement, commercialization and profits,
other techniques were developed for introducing foreign
0167-7799/00/$ – see front matter © 2000 Elsevier Science Ltd. All rights reserved. PII: S0167-7799(00)01444-X
257
TTEC June 2000
4/5/00
11:15 am
Page 258
REVIEWS
Table 1. Comparison of the nuclear and plastid genomes of angiosperms
Nuclear genome
Plastid genome
Chromosomes
Two copies of each of many chromosomes;
the number of chromosomes per diploid
cell is species-specific
~60 copies of a single circular
chromosome per plastid
~50–60 chloroplasts per cell
Genes per chromosome
Could be thousands
~120–150
Arrangement and
transcription of genes
Each gene is separate and is transcribed
individually
Many genes are in operons and are
transcribed together
DNA into cells. Among these techniques, microprojectile
bombardment7 has been broadly effective and has been
particularly successful for producing transgenic maize8,9.
The rapid increase, over the past few years, in the
areas planted with Ti-plasmid-transformed cotton and
microprojectile-transformed maize attests to the success
of the methods for transforming plant nuclei. However,
the current techniques have important limitations. One
significant problem, which is at best an inconvenience
and at worst a blocking impediment, is that the expression level of an introduced gene can vary greatly from
one transformed plant to another. Consequently, many
transformants with the same gene usually need to be
produced and tested to identify a strongly expressing
plant strain suitable for further research or for commercialization. This effect is thought to be attributable,
at least in part, to poorly understood ‘position effects’.
The notion is that the location at which a gene is
inserted on a chromosome – and on which chromosome – affects how strongly it is expressed. Methods for
directing the insertion of foreign genes to specific sites
in the nuclear genomes of plants have not been developed but a valuable method for creating small mutations
in known genes in situ has been described recently10,11.
Multiple genes
The consequences of variable expression are compounded if more than one gene needs to be introduced.
Although some useful traits result from the presence of
a single gene (e.g. the production of Bacillus thuringiensis
toxin to protect against insects, of a viral coat protein
to protect against viruses or of an enzyme to inactivate
a herbicide), the vast majority of agronomic traits are
multigenic, that is, they result from the action of
several genes. A biosynthetic pathway is perhaps the
simplest example: a series of enzymes, each encoded by
one or more genes, is needed to yield a product.
Natural examples include the biosynthetic chains that
produce cell-wall components, chlorophyll, amino
acids, and so on. Indeed, most of the characters that
concern plant breeders are quantitative traits; they are
controlled polygenetically. As discussed above, introducing a single gene usually yields some transformants
that express the new gene strongly and others that
express it poorly. The best expressing strain must be
selected from among these. Suppose that this gene
encodes one of several enzymes in a biosynthetic chain
that is to be introduced into a plant. After selecting
transformants that express the first gene well, the best
of these would need to be transformed with a gene for
an enzyme that catalyses the next step in the biosynthetic
258
chain, followed by another round of selections from
among many transformants. Obviously, the more
enzymes in the chain, the more times these steps must
be repeated.
However, in an exciting and skilful seven-year-long
effort12, Potrykus et al. recently introduced into rice a
set of three genes for a short biosynthetic chain that
resulted in ‘yellow rice’. (They were fortunate in being
able to introduce two of these genes at one time.) This
chain of genes produced the enzymes required to
enable rice grains to form b-carotene, which humans
convert into vitamin A13. The complications attendant
on introducing a more extended biosynthetic chain and
on attaining other major plant biotechnology goals are
obvious. Two separate genes introduced into a plant in
a single T-DNA fragment may be expressed to different relative extents in one transformant than in another.
Thus, contrary to what one might have thought, cloning
several genes into a single T-DNA does not generally
avoid the compounded, variable expression problem.
An interesting single-gene approach for simultaneously introducing several new proteins into a cell via
nuclear gene transformation was taken by Dasgupta et
al.14 They constructed and introduced a large single
gene encoding a polyprotein containing three enzymes
into a plant. The enzyme proteins were separated from
one another by the amino acid sequence recognized by
the tobacco-vein-mottling-virus N1a proteinase, which
was one of the three enzymes in the polyprotein. The
three enzymes were detected as individual active peptides in the transformant, showing that the N1a proteinase cut the individual enzymes out of the polyprotein
and that all three enzymes were functional.
In nature, however, the problem is approached in
various other ways. Operons (groups of genes with a
single promoter), which often encode enzymes of the
same biosynthetic chain, are common in bacteria. The
single transcript of an entire multigenic operon carries
information required for the production of several proteins. Nuclear genes are transcribed singly; they are not
arranged in operons. However, chloroplast genes are
often present in operons. Some pertinent features of
nuclear and plastid genomes are compared in Table 1.
Chloroplasts for genetic engineering
Plant cells have three gene-containing compartments:
the nuclear–cytoplasmic system, mitochondria and
plastids. There are several types of plastid including:
(1) chlorophyll-containing chloroplasts; (2) yellow,
orange or red carotenoid-containing chromoplasts;
(3) starch-storing amyloplasts; (4) oil-containing elaioplasts;
TIBTECH JUNE 2000 (Vol. 18)
TTEC June 2000
4/5/00
11:15 am
Page 259
REVIEWS
(5) proplastids (plastid precursors found in most plant
cells); and (6) etioplasts (partially developed chloroplasts
that form in dark-grown seedlings). The conversion of
photosynthetic chloroplasts into yellow carotenoid-rich
chromoplasts is seen in the ripening of bananas; the
conversion of chloroplasts to lycopene-containing red
chromoplasts is seen in the ripening of tomatoes. Each
compartment of the eukaryotic cell is unique. A particular biochemistry can be favored in one compartment
(e.g. chloroplasts or chromoplasts) while the environment in another compartment (e.g. the cytoplasm) is
unfavorable.
The photosynthetic apparatus of the chloroplast
captures light energy and converts it into biologically
useful energy that is, for example, available for: (1) manufacturing sugars and amino acids; (2) reducing nitrate
to ammonium; (3) making starch; (4) synthesizing complex organic compounds; and (5) assembling chloroplast proteins that are required for the chloroplasts
themselves to function. (Non-green plastids use energy
produced elsewhere in the cell for metabolism.) It is
easy to imagine modifying or adding to existing
biosynthetic chains, which are already linked to the
energy-converting machinery of the chloroplast, to
produce and accumulate desirable new products. The
chloroplast genomes of flowering plants encode perhaps 150 genes; the remainder of the hundreds or thousands of plastid proteins are the products of nuclear
genes. The compositions of chloroplasts and other plastids can be modified by inserting genes into the nuclear
genome; a chimeric gene can be constructed with an
N-terminal sequence that targets its protein product to
the chloroplast.
Why plastids?
However, why do we need to send a protein into the
plastid? Some syntheses other than photosynthesis are
only carried out in plastids, probably because one or
another feature of the organelle’s environment does not
exist elsewhere in the cell. Furthermore, the cytoplasm
and chloroplast contain different proteases, and a protein might survive better in one compartment than in
the other. A specific type of plastid might be a good
place to accumulate certain proteins or their biosynthetic products that would be harmful if they were
present in large amounts in the cytoplasm or in a plastid
of a different type. The various forms of plastid (amyloplasts, chromoplasts, etc.) have desirable properties as
places to conduct reactions and to accumulate proteins
or products of enzymes, and they can be exploited
using known nuclear gene transformation methods.
Nevertheless, the limitations of plant nuclear transformation technology discussed above remain. Fortunately,
there is an alternative: new genes can be introduced
directly into the plastid genome. This has several advantages, although some technical problems have still to
be solved to simplify the procedure.
Methods for transforming plastids
In 1988, it was reported that a deletion in the chloroplast gene atpB that rendered the green alga Chlamydomonas reinhardtii incapable of carrying out photosynthesis could be corrected by introducing the wild-type
gene15. The DNA was deposited on tungsten microprojectiles ~1 mm in size and propelled into cells spread
TIBTECH JUNE 2000 (Vol. 18)
on an agar plate using a gunpowder charge7. Later, compressed gases were used for biolistic transformations16,17.
The DNA became incorporated into the Chlamydomonas
chloroplast chromosome by homologous recombination.
Thus, unlike nuclear transformation, DNA can be
directed to a specific site within the plastid chromosome. Blowers et al. showed that foreign DNA bordered by chloroplast DNA sequences is incorporated
and stably maintained in the Chlamydomonas chloroplast
chromosome18.
Daniell et al.19 and Svab et al.20 took advantage of this
knowledge and introduced DNA into tobacco chloroplasts in cultured cells in situ or in leaf tissue. Transient
expression was observed19 and, in addition, transformed
cells were regenerated into plants with modified chloroplasts20. The usefulness of plastid transformation for
biotechnology was shown strikingly by McBride et al.,
who introduced the gene encoding the B. thuringiensis
lepidopteran protoxin into the chloroplast chromosomes of tobacco plants using the microprojectile
method. In the resulting plants, this protoxin constituted 2–3% of the soluble leaf protein21. This very high
level of transgene expression was attributed to the presence of ~50 chloroplasts per leaf cell and 60–100 chromosomes per plastid. Each leaf cell contains ~3000
copies of each gene, present once in each plastid
chromosome. However, the large number of copies of
the gene in the cell is unlikely to be the entire explanation: some chloroplast genes encode proteins that are
amongst the most abundant in the cell, but others are
much less highly expressed. Also, the mRNAs for some
nuclear genes that are present in <10 copies per cell
are as abundant as the most abundant plastid-encoded
mRNAs.
Two additional methods for delivering DNA into
plastids have recently been developed: a polyethylene
glycol-mediated method and direct injection of DNA
in situ. Nicotiana plumaginifolia22 and Nicotiana tabacum23
with transformed chloroplasts have been produced by
exposing protoplasts to polyethylene glycol and
reporter DNA bordered by chloroplast sequences to
enable homologous recombination to occur. The polyethylene-glycol method holds out the promise of the
capacity to generate more cells with transformed plastids more readily than by the biolistic procedure. However, the limiting step in this case is regenerating plants
from protoplasts. For example, immature embryos,
rather than protoplasts, are used as the starting material
for successfully producing maize and rice varieties with
nuclear transgenes. Although they have not yet regenerated any plants, Knoblauch et al. have shown that DNA
is expressed after it is injected directly into individual
chloroplasts in photosynthetic leaf cells of tobacco24.
Operons
Many chloroplast genes are grouped in operons.
Multicistronic mRNAs transcribed from these operons
can be processed to produce mRNAs for single proteins. Introducing blocks of foreign genes in a single
operon would avoid the complications inherent in
putting one gene at a time into random locations in the
nuclear genome. Introducing genes in a single block
would be the easiest but, because the position on the
plastid chromosome is not known to affect the rate of
transcription of a gene, there might be no disadvantage
259
TTEC June 2000
4/5/00
11:15 am
Page 260
REVIEWS
to having transgenes with the same kind of promoter
at separate locations on the chromosome.
Because insertion is to a specified site on the chromosome, if particular sites are found to affect the transcription of an inserted gene (e.g. through effects on
DNA conformation that affect transcription25), the site
of insertion of transgenes could be selected in order to
take advantage of these effects. This is better than being
at the mercy of undirected insertion and position
effects. (Of course, uniform transcription does not assure
uniform stability or equal translation of transcripts. But
these are separate problems and are faced in whichever
cellular compartment the gene is transcribed.)
The expression of some plastid genes, as is the case
for some nuclear genes, is regulated by environmental
factors (e.g. light) and other genes are expressed only
in specific types of plastid and cell. Such specificities
can be taken advantage of in genetic engineering.
There is a lot of information available about plastidgene DNA-dependent RNA polymerases, promoters
and enhancers, and all of this knowledge can be exploited
for genetically engineering the plant’s relatively simple
plastid genome.
Inheritance
The chloroplasts of angiosperms are generally transmitted only by the maternal parent. This means that
chloroplast genes are not present in pollen. Consequently, a trait introduced by genetically engineering
the chloroplast genome would not be unintentionally
transferred to sexually compatible relatives of the crops
that might be growing nearby26. Furthermore, the
plastid-chromosome-encoded protein would not be
produced in the pollen and thus would not affect insects
that feed on pollen or pollen-coated plant tissues.
Pros, cons and uncertainties
Plastids are important alternatives to the cytoplasm as
sites in which to engineer enzymatic reactions or to
make, process or store proteins or enzyme products.
The internal environments of some differentiated forms
of plastids (e.g. chloroplasts, chromoplasts and amyloplasts) might be more favorable than those of the cytoplasm for certain enzymatic reactions, for modifying or
processing a particular protein and so on. A protein
might be more stable in the plastid than in the cytoplasm because of differences in proteases, ions and other
constituents. It might also be advantageous to place
energy-intensive biochemical reactions close to the site
of photosynthesis. In addition, if the protein that accumulates or if the product of an enzymatic reaction is
deleterious to the cell if present in the cytoplasm, it
might not be harmful in some form of plastid. However, the plastid could be the wrong place for a process
if the product must leave the plastid and go to another
place in the plant. Various small organic molecules are
transported out of chloroplasts but very little is currently known about the export of many molecules from
this organelle.
Proteins made in the cytoplasm can be directed into
the plastids so why consider putting genes directly into
the plastid genome? First, the level of expression might
be more predictable. Second, a group of genes can be
introduced as an operon for transcription as a unit in
the plastid or as a series of independently transcribed
260
DNA sequences, all in a single transformation. Third,
foreign genes will not disperse to related species
growing within pollination range of the crop because
chloroplast genes are not present in most angiosperm
pollens. In addition, locating genes in plastids might
offer different possible expression patterns in the plant
than nuclear localization (e.g. in all cells containing
proplastids or chloroplasts or amyloplasts).
The present range for plastid transformation
Chloroplast transformation has been very successful
for Chlamydomonas reinhardtii and tobacco. This has
allowed important experiments to be done on the molecular biology of chloroplasts and on photosynthetic
processes. Among higher plants, fertile plants with
genetically modified chloroplast chromosomes have
been reported, to date, only in tobacco. Low-efficiency
plastid transformation has been reported for Arabidopsis
thaliana27 but none of the transformed plants that regenerated were fertile. Chloroplasts in potato-leaf cells have
been transformed with the bacterial gene for spectinomycin and streptomycin resistance (aadA) as well as with
the green fluorescent protein (GFP) gene of the jellyfish Aequorea victoria28. Plastids in suspension cultures of
embryogenic rice cells have been transformed with an
aadA–GFP fusion gene29. In these cases, DNA was delivered on microprojectiles and transgenic plants were
recovered, but in neither case were fertile plants recovered.
However, these results are encouraging, especially
when combined with the observation that GFPencoding DNA attached to a plastid rRNA promoter
(delivered on microprojectiles) is transiently expressed
in non-green starch-storing amyloplasts in potato tuber
tissue and in chromoplasts in marigold petals, carrot
roots and pepper fruits30. First, these results show that
foreign genes can be expressed in non-green plastids
and, at least in the case of rice, that the DNA can be
incorporated into the genome of non-green plastids.
(This suggests that the range of target tissues for plastid
transformation extends beyond chloroplast-containing
cells.) Second, the results with potato and rice provide
a good start for the broadening of the species range of
the technology. These limited successes are heartening
but a great deal remains to be done to extend the
usefulness of chloroplast transformation technology to
these and other economically important crops.
Heterogeneity, heteroplasmy and stability
Although the inability to regenerate fertile, transgenic, economically important plants readily is a major
impediment (just as it was in the early days of nuclear
transformation), there is every reason to expect that this
difficulty will be overcome with one species at a time
in chloroplast genetic engineering, as it has been and
continues to be in plant nuclear biotechnology.
However, securing genetically stable lines of plants
with transgenic plastids presents a technically challenging novel problem. Each cell can have 50, 75 or more
plastids and each chloroplast has ~60 chromosomes,
which means that there are 3000–4000 chromosomes
in about 60 compartments in each cell. How can we
obtain homoplasmic plants in which every plastid
chromosome carries the inserted transgene? Or, at least,
how can we ensure that the plants do not lose all the
transgene-containing chromosomes?
TIBTECH JUNE 2000 (Vol. 18)
TTEC June 2000
4/5/00
11:15 am
Page 261
REVIEWS
Chlamydomonas, by contrast, is a single-celled flagellate, each of whose cells contains a single chloroplast
with ~60 chromosomes, and homoplasmic chloroplast
transformants can generally be isolated after continued
culture under selection. A wild-type plastid chromosome
in a transplastomic crop plant could persist under the
‘protection’ of the product of the introduced gene
present in other chromosomes in the same or other
plastids and take over in the absence of selection for the
transgene.
The current major approach to obtaining homoplasmic plants is repeated selection from transformed tissues in culture, when possible, followed by regenerating
plants. Unfortunately, repeated cell culturing puts a
further restraint on the kinds of plant material that can
be used for regeneration. A different strategy might
be to make the transgene (or the transgene product
fused to another protein coding sequence) essential
for cell survival, thus providing constant selection and
rendering the presence of wild-type chromosomes
irrelevant. Another course might be temporarily (i.e.
reversibly) reducing the number of chloroplasts per cell
and the number of chromosomes per plastid. Antisense
suppression of one of the genes controlling plastid division in A. thaliana results in the number of chloroplasts
per leaf cell dropping from 100 to one31, and Chlamydomonas chloroplast, but not nuclear, DNA is temporarily reduced to one-seventh its normal level (approximately 10 rather than 60 chromosomes per plastid) after
6–7 doublings in the presence of the thymidylatesynthase inhibitor fluorodeoxyuridine32. However,
thymine starvation-induced mutagenesis occurs in bacteria, T4 bacteriophage and, probably, Chlamydomonas
chloroplast DNA33, which means that this might be a
clue to how to reduce the chloroplast DNA rather than
a literal solution.
It remains to be seen whether one or both of these
sorts of approaches can be used to address the vexing
problem of wild-type chromosomes lurking in the
background but ready to ‘take over’. Concerted efforts
are likely to overcome the present limitations but technology that can easily solve or circumvent the problem
has not yet been developed. It was recognized many
years ago that chloroplasts in the giant alga Acetabularia
might not be entirely independent of one another and,
indeed, might be a single fusing and dividing mass34.
The recent observation of connections between chloroplasts in cells of higher plants35 is a striking reminder of
how little we know about plastid biology. As we learn
about the subject, we are likely to acquire new outlooks
on how to approach the heteroplasmy problem directly
and effectively. Some aspects of nuclear and plastid
genome transformation are compared in Table 2.
New and renewed possibilities?
The development of multiple sites for genetically
engineering plant cells raises new prospects for genetic
engineering. It also prompts renewed consideration of
two examples of major modifications of plant metabolism involving plastids that could be very important
for agriculture: (1) enabling plants to fix N2 and
(2) improving photosynthetic CO2 fixation. Although
both possibilities might, in reality, still be a long way
off, there are reasons to be hopeful. First, we continue
to learn more about how to take advantage of multiple
sites in the cell for introducing foreign genes and about
how to control plant gene expression, and (particularly
through proteomics) we should come to know more
about the kinds of reactions favored in one cell compartment over others. Second, at the same time, we are
becoming increasingly better-informed about the biochemistry and physiology of N2 fixation and natural
adaptations for increasing CO2 fixation.
Nitrogen fixation
It was first proposed approximately 20 years ago that
plastids might be suitable and appropriate sites for engineering into plants the capacity for nitrogen fixation36,37.
Nitrogen fixation requires large amounts of energy in
the form of ATP and chloroplasts make this as a product of photosynthetic energy transduction. The sensitivity of the enzyme nitrogenase to O2, a major product of photosynthesis, would appear to preclude N2
fixation by chloroplasts but cyanobacteria also conduct
oxygenic photosynthesis despite fixing N2 very effectively. Some cyanobacteria separate the two processes
Table 2. Introducing genes into nuclear and plastid genomes
Nuclear genome
Plastid genome
Insertion of foreign DNA
into genome
Undirected; multiple insertions are common.
Directed to a specific site by homologous recombination.
Transcription of
introduced genes
Affected by the promoter, the type of cell
and the site of insertion of the gene into
the genome. Each introduced gene is
expressed individually.
Affected by the promoter, the type of plastid and the
type of cell. The location on the chromosome is not
known to affect transcription. A set of genes could be
introduced as an operon or as individual transcription
units.
Current limitations
The level of expression of an introduced
gene is unpredictable.
Each gene in a set required for a new
multigenic trait or biosynthetic pathway
may have to be introduced separately
and sequentially. The level of expression
of each separately introduced gene is
unpredictable.
Obtaining homoplasmic transformed strains can be
difficult: the development of easier new methods will be
important. Alternatively, methods are needed to retain
transgene(s) in the presence of untransformed plastid
chromosomes.
To date, tobacco is the only crop in which fertile plants
with plastid transgenes have been described. Reports
on other crops are promising.
TIBTECH JUNE 2000 (Vol. 18)
261
TTEC June 2000
4/5/00
11:15 am
Page 262
REVIEWS
Box 1. Additional considerations in
targeting genes, biosynthetic paths and
enzyme products
• The stability and accumulation of a particular RNA
or protein may be greater or lesser in the nuclearcytoplasmic compartment than in a plastid compartment or vice versa.
• Conditions for a particular enzymatic reaction may be
more or less favorable in the nuclear-cytoplasmic compartment than in a plastid compartment or vice versa.
• More of the product(s) of an enzymatic reaction or of
a chain of reactions may accumulate and not affect
the cell in the nuclear-cytoplasmic compartment than
in a plastid compartment or vice versa.
• If gene containment is important, the maternal transmission of plastids may be advantageous.
• A single type of plastid does not necessarily occur in
only a single type of cell. Therefore, a different pattern
of gene expression may be achieved by inserting
genes into the plastid genome than in the nuclear
genome.
temporally and others produce heterocysts for N2 fixation. Heterocysts have thickened walls that impede
the passage of O2 but, of paramount importance, the
heterocyst’s photosynthetic apparatus lacks the capacity
to evolve oxygen but retains the ability to produce ATP.
CO2 fixation in C3 plants
In C3 plants (including most crop species), CO2 is
fixed to the five-carbon compound ribulose-1,5-bisphosphate. Two molecules of 3-phosphoglycerate
(a C3 compound) are produced from the product of
this fixation. The phosphoglycerate is reduced to 3phosphoglyceraldehyde using the ATP and reduced
nicotinamide adenine dinucleotide phosphate (NADPH)
that are produced photosynthetically. All chloroplasts
in all the cells of a C3 leaf contain ribulose bisphosphate carboxylase–oxygenase (Rubisco), the enzyme
that catalyses this fixation of CO2. However, as well as
fixing CO2 to ribulose-1,5-bisphosphate to produce
two molecules of phosphoglycerate, Rubisco can catalyse the oxygenation of the same five-carbon compound
using O2 to produce one molecule of phosphoglycerate and one molecule of phosphoglycolate; O2 and
CO2 compete for the active site on Rubisco. As the rate
of photosynthesis increases (e.g. as the light intensity
increases), so does O2 production, resulting in more
competition with CO2 for Rubisco. As a result, C3
photosynthesis saturates at about one third of the
maximum light intensity under which many crop plants
grow.
CO2 fixation in C4 plants
C4 plants can take advantage of higher light intensities. In C4 plants, such as maize and sorghum, there
are two types of photosynthetic leaf cell: bundle-sheath
cells, which surround the vascular bundles, and mesophyll cells, which occupy the space between the cylinders of bundle-sheath cells and the upper and lower epidermis of leaves. Bundle-sheath cells contain Rubisco;
mesophyll cells do not. Carbon dioxide is fixed first in
mesophyll cells by addition to phosphoenolpyruvate to
form oxaloacetate (a C4 compound). Then, in maize,
262
for example, oxaloacetate produced in mesophyll cells is
reduced to malate, which is transferred to the adjacent
bundle-sheath cells. In the bundle-sheath cells, CO2 is
released and refixed by Rubisco in a low-O2 environment,
because bundle-sheath chloroplasts, like cyanobacterial
heterocysts, have little or no oxygen-evolving capacity.
Improving CO2 fixation
Could C3 plants be converted into C4 plants? C4
plants are better adapted to more arid conditions and
to using higher light intensities than C3 plants. New
information about the molecular basis for suppressing
Rubisco production in the mesophyll cells of maize38
is one step towards making this possible. Of course, it
would also be necessary to selectively eliminate O2 production in chloroplasts with Rubisco and to increase
the activities of phosphoenolpyruvate carboxylase and
other enzymes involved in the fixation of CO2 into
oxaloacetate as in Rubisco-free mesophyll cells of C4
plants. An alternative and simpler way to improve the
effectiveness of CO2 fixation by Rubisco might be by
genetically installing a CO2-concentrating mechanism39.
Important knowledge has been gained about the
genes and enzymes of nitrogen fixation since the first
suggestions were made that bundle-sheath plastids and
cyanobacterial heterocysts were excellent models and
that bundle-sheath plastids could be good sites for
engineering N2 fixation36; this has been the subject of
very thoughtful discussion by Dixon et al.40
Future prospects
Relatively few genes have been introduced into
plants for commercial purposes. The importance of the
availability of separate, metabolically unique compartments in cells will become increasingly evident as more
diverse metabolic products and enzymatic reactions are
introduced into plants. The possibility that a non-cytoplasmic environment, such as the mitochondrion or a
particular form of plastid, will favor a particular enzymatic reaction and/or the accumulation of a protein or
metabolic product will be considered with increasing
frequency. (Tables 1 and 2, and Box 1, summarize the
features of the nuclear and plastid genomes that are pertinent to genetic transformation and recapitulate other
considerations for judging whether a gene, a biosynthetic chain or an enzyme product is best targeted to
plastids or to another cellular compartment.)
The major change in plant biotechnology, however,
will come with the need to introduce sets of genes in
order to express traits determined by multiple genes. In
these cases, especially if large amounts of energy are
needed to drive one or more of the reactions catalysed
by enzymes encoded by newly introduced genes, the
method of first choice is likely to be its introduction
into the chloroplast genome. Indeed, even if the processes
are not energy intensive, it might be easier to introduce
multiple genes into the plastid genome than into the
nuclear genome.
Current knowledge has broadened the possibilities
for genetically engineering plants. The nuclear–cytoplasmic system might be the most favorable place in a
cell for some new biosynthetic processes but other locations in the cell, among them plastids, might be better
for others. As for accumulating metabolic products or
proteins, the mantra is: the product might be safer in
TIBTECH JUNE 2000 (Vol. 18)
TTEC June 2000
4/5/00
11:15 am
Page 263
REVIEWS
the plastid and the cell might be safer with the product
in the plastid.
Acknowledgments
The work in the author’s laboratory and the preparation of this paper were supported in part by research
grants from the National Institute of General Medical
Sciences of the National Institutes of Health of the
USA.
References
1 Van Larebeke, N. et al. (1974) Large plasmid in Agrobacterium
tumefaciens essential for crown gall-inducing ability. Nature 252,
169–170
2 Gordon, M.P. (1998) Discovery of the T-DNA of Agrobacterium
tumefacieus. In Discoveries in Plant Biology (Vol. 1) (Kung, S.D. and
Yang, S.F., eds), pp. 111–114, World Scientific Publishing
3 Xu, Z-H. (1998) The discovery of plant tissue culture. In Discoveries in Plant Biology (Vol. 2) (Kung, S.D. and Yang, S.F., eds), pp.
287–316, World Scientific Publishing
4 Hiei, Y. et al. (1997) Transformation of rice mediated by Agrobacterium
tumefaciens. Plant Mol. Biol. 35, 205–218
5 Ishida, Y. et al. (1996) High efficiency transformation of maize
(Zea mays L.) mediated by Agrobacterium tumefaciens. Nat. Biotechnol.
14, 745–750
6 Cheng, M. et al. (1997) Genetic transformation of wheat mediated
by Agrobacterium tumefaciens. Plant Physiol. 115, 971–980
7 Klein, T.M. et al. (1987) High velocity microprojectiles for delivering nucleic acids into living cells. Nature 327, 70–73
8 Fromm, M.E. et al. (1990) Inheritance and expression of chimeric
genes in the progeny of transgenic maize plants. Biotechnology 8,
833–839
9 Gordon-Kamm, W.J. et al. (1990) Transformation of maize cells and
regeneration of fertile transgenic plants. Plant Cell 2, 603–618
10 Zhu, T. et al. (1999) Targeted manipulation of maize genes in vivo
using chimeric RNA–DNA oligonucleotides. Proc. Natl. Acad. Sci.
U. S. A. 96, 8768–8773
11 Beetham, P.R. et al. (1999) A tool for functional plant genomics:
chimeric RNA–DNA oligonucleotides in vivo gene-specific
mutations. Proc. Natl. Acad. Sci. U. S. A. 96, 8774–8778
12 Gura, T. (1999) New genes boost rice nutrients. Science 285,
994–995
13 Ye, X. et al. (2000) Engineering the provitamin A (b-carotene)
biosynthetic pathway into (carotenoid free) rice endosperm. Science
287, 303–305
14 Dasgupta, S. et al. (1998) Co-ordinated expression of multiple
enzymes in different subcellular compartments in plants. Plant J. 16,
107–116
15 Boynton, J.E. et al. (1988) Chloroplast transformation in Chlamydomonas with high velocity microprojectiles. Science 240, 1534–1538
16 Sanford, J.C. et al. (1991) An improved helium-driven biolistic
device. Technique 3, 3–16
17 Finer, J.J. et al. (1992) Development of the particle inflow gun for
DNA delivery to plant cells. Plant Cell Rep. 11, 323–328
18 Blowers, A.D. et al. (1989) Studies on Chlamydomonas chloroplast
transformation: foreign DNA can be stably maintained in the
chromosome. Plant Cell 1, 123–132
19 Daniell, H. et al. (1990) Transient foreign gene expression in chloroplasts of cultured tobacco cells after biolistic delivery of chloroplast
vectors. Proc. Natl. Acad. Sci. U. S. A. 87, 88–92
20 Svab, Z. et al. (1990) Stable transformation of plastids in higher plants.
Proc. Natl. Acad. Sci. U. S. A. 87, 8526–8530
21 McBride, K.E. et al. (1995) Amplification of a chimeric Bacillus gene
in chloroplasts leads to an extraordinary level of an insecticidal
protein in tobacco. Biotechnology 13, 362–365
22 O’Neill, C. et al. (1993) Chloroplast transformation in plants: polyethylene glycol (PEG) treatment of protoplasts is an alternative to
biolistic delivery systems. Plant J. 3, 729–738
23 Koop, H.U. et al. (1996) Integration of foreign sequences into the
tobacco plastome via polyethylene glycol-mediated protoplast transformation. Planta 199, 193–201
24 Knoblauch, M. et al. (1999) A galinstan expansion femtosyringe for
microinjection of eukaryotic organelles and prokaryotes. Nat.
Biotechnol. 17, 906–909
25 Stirdivant, S.M. et al. (1985) DNA supercoiling affects in vitro transcription of two maize chloroplast genes differently. Proc. Natl. Acad.
Sci. U. S. A. 82, 4886–4890
26 Daniell, H. et al. (1998) Containment of herbicide resistance through
genetic engineering of the chloroplast genome. Nat. Biotechnol. 16,
345–348
27 Sikdar, S.R. et al. (1998) Plastid transformation of Arabidopsis thaliana.
Plant Cell Rep. 18, 20–24
28 Siderov, V. et al. (1999) Stable chloroplast transformation in potato:
use of green fluorescent protein as a plastid marker. Plant J. 19, 209–216
29 Khan, M.S. et al. (1999) Fluorescent antibiotic resistance marker for
tracking plastid transformation in higher plants. Nat. Biotechnol. 17,
910–915
30 Hibbard, J.M. et al. (1998) Transient expression of green fluorescent
protein in various plastid types following microprojectile bombardment. Plant J. 16, 627–632
31 Osteryoung, K.W. et al. (1998) Chloroplast divisions in higher plants
requires members of two functionally divergent gene families with
homology to bacterial ftZ. Plant Cell 10, 1991–2004
32 Wurtz, E.A. et al. (1977) Pertubation of chloroplast DNA amounts
and chloroplast gene transmission in Chlamydomonas reinhardtii by
fluorodeoxyuridine. Proc. Natl. Acad. Sci. U. S. A. 74, 4552–4556
33 Wurtz, E.A. et al. (1979) A specific increase in chloroplast gene
mutations following growth of Chlamydomonas in 5-fluorodeoxyuridine. Mol. Gen. Genet. 170, 235–242
34 Woodcock, C.L.F. et al. (1970) Evidence for variation in the quantity of DNA among plastids of Acetabularia. J. Cell Biol. 44, 361–375
35 Kohler, R.H. et al. (1997) Exchange of protein molecules through
connections between higher plant plastids. Science 276, 2039–2042
36 Bogorad, L. (1979) The chloroplast, its genome and possibilities
for genetically manipulating plants. In Genetic Engineering (Vol. 1)
(Setlow, J.K. and Hollaender, A., eds), pp. 181–203, Plenum Press
37 Merrick, M. and Dixon, R. (1984) Why don’t plants fix nitrogen?
Trends Biotechnol. 2, 162–166
38 Purcell, M. et al. (1995) Red/far-red and blue light-responsive
regions of maize rbcS-m3 are active in bundle sheath and mesophyll
cells, respectively. Proc. Natl. Acad. Sci. U. S. A. 92, 11504–11508
39 Reiskind, J.B. et al. (1997) Evidence that inducible C4-type photosynthesis is a chloroplastic CO2-concentrating mechanism in Hydrilla,
a submersed monocot. Plant Cell Environ. 20, 211–220
40 Dixon, R. et al. (1997) Nif gene transfer and expression in chloroplasts: prospects and problems. Plant Soil 194, 193–203
Reference lists in TIBTECH articles
The lists of references accompanying articles in TIBTECH are intended to provide readers with a general
indication of where further information on the topic can be found. They are not complete lists of every
paper published on the topic.
Authors writing for TIBTECH are given a strict limit as to the number of references they may include, although
this limit may vary for different types of articles.
TIBTECH JUNE 2000 (Vol. 18)
263