Mutations in a plastid-localized elongation factor G alter early stages of plastid development in Arabidopsis thaliana.

Abstract

Proper development of plastids in embryo and seedling tissues is critical for plant development. During germination, plastids develop to perform many critical functions that are necessary to establish the seedling for further growth. A growing body of work has demonstrated that components of the plastid transcription and translation machinery must be present and functional to establish the organelle upon germination.
We have identified Arabidopsis thaliana mutants in a gene that encodes a plastid-targeted elongation factor G (SCO1) that is essential for plastid development during embryogenesis since two T-DNA insertion mutations in the coding sequence (sco1-2 and sco1-3) result in an embryo-lethal phenotype. In addition, a point mutation allele (sco1-1) and an allele with a T-DNA insertion in the promoter (sco1-4) of SCO1 display conditional seedling-lethal phenotypes. Seedlings of these alleles exhibit cotyledon and hypocotyl albinism due to improper chloroplast development, and normally die shortly after germination. However, when germinated on media supplemented with sucrose, the mutant plants can produce photosynthetically-active green leaves from the apical meristem.
The developmental stage-specific phenotype of the conditional-lethal sco1 alleles reveals differences in chloroplast formation during seedling germination compared to chloroplast differentiation in cells derived from the shoot apical meristem. Our identification of embryo-lethal mutant alleles in the Arabidopsis elongation factor G indicates that SCO1 is essential for plant growth, consistent with its predicted role in chloroplast protein translation.

Full text

Page 1

BioMed Central
Page 1 of 10
(page number not for citation purposes)
BMC Plant Biology
Open Access
Research article
Mutations in a plastid-localized elongation factor G alter early
stages of plastid development in Arabidopsis thaliana
Nicholas J Ruppel and Roger P Hangarter*
Address: Department of Biology, Indiana University, Bloomington, IN, 47405, USA
Email: Nicholas J Ruppel - nruppel@indiana.edu; Roger P Hangarter* - rhangart@indiana.edu
* Corresponding author
Abstract
Background: Proper development of plastids in embryo and seedling tissues is critical for plant
development. During germination, plastids develop to perform many critical functions that are
necessary to establish the seedling for further growth. A growing body of work has demonstrated
that components of the plastid transcription and translation machinery must be present and
functional to establish the organelle upon germination.
Results: We have identified Arabidopsis thaliana mutants in a gene that encodes a plastid-targeted
elongation factor G (SCO1) that is essential for plastid development during embryogenesis since
two T-DNA insertion mutations in the coding sequence (sco1-2 and sco1-3) result in an embryo-
lethal phenotype. In addition, a point mutation allele (sco1-1) and an allele with a T-DNA insertion
in the promoter (sco1-4) of SCO1 display conditional seedling-lethal phenotypes. Seedlings of these
alleles exhibit cotyledon and hypocotyl albinism due to improper chloroplast development, and
normally die shortly after germination. However, when germinated on media supplemented with
sucrose, the mutant plants can produce photosynthetically-active green leaves from the apical
meristem.
Conclusion: The developmental stage-specific phenotype of the conditional-lethal sco1 alleles
reveals differences in chloroplast formation during seedling germination compared to chloroplast
differentiation in cells derived from the shoot apical meristem. Our identification of embryo-lethal
mutant alleles in the Arabidopsis elongation factor G indicates that SCO1 is essential for plant
growth, consistent with its predicted role in chloroplast protein translation.
Background
In oilseed plants such as Arabidopsis (Arabidopsis thaliana)
and rapeseed (Brassica napus), developing embryos are
green and cells in these embryos develop functional chlo-
roplasts [1]. The green embryos are capable of photosyn-
thesis and have been shown to fix carbon crucial to the
biosynthesis of seed storage oils [2-4]. In experiments
with cultured rapeseed embryos and siliques, light was
found to increase embryo growth-rates, which correlated
both with improved carbon sequestration and with its uti-
lization in seed oil synthesis [5]. These effects were largely
negated by inhibition of photosynthesis, and their studies
indicated that it is the reductant and/or ATP produced by
photosynthesis in green embryos that is important for
normal embryo growth and seed development.
Arabidopsis embryos begin to develop chloroplasts and
appear green around 5 days after pollination and the chlo-
Published: 13 July 2007
BMC Plant Biology 2007, 7:37 doi:10.1186/1471-2229-7-37
Received: 1 December 2006
Accepted: 13 July 2007
This article is available from: http://www.biomedcentral.com/1471-2229/7/37
© 2007 Ruppel and Hangarter; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Page 2

BMC Plant Biology 2007, 7:37http://www.biomedcentral.com/1471-2229/7/37
Page 2 of 10
(page number not for citation purposes)
roplasts remain present for up to approximately 12 days
after pollination [6]. When seeds are maturing during late
embryogenesis (>12 days), the chloroplasts dedifferenti-
ate and lose their chlorophyll, starch, and internal mem-
branes to seed storage reserves, which results in the
formation of white embryos in mature Arabidopsis seed
[6]. In soybean (Glycine max), a small simple plastid called
an eoplast has been found in fully mature embryo cells
[7]. Eoplasts resemble proplastids but are derived from
chloroplasts. Thus far, eoplasts have not been observed in
mature Arabidopsis embryos, largely because the embryo
cells are so densely packed with lipid and protein bodies
[8]. Nevertheless, a basal-state plastid must be maintained
in the cells of fully mature Arabidopsis embryos since
chloroplasts, amyloplasts, and the various other plastid
types re-develop upon seedling germination. The develop-
ment of these plastids early after germination can be criti-
cal for seedling survival since, in addition to
photosynthesis and the production of starch, plastids are
also involved in the biosynthesis of fatty acids [9], nucleic
acids, and amino acids [10].
Although a good deal is known about the physiological
and biochemical functions of chloroplasts during embryo
growth and seed production, investigations of chloroplast
development during embryogenesis have been largely
descriptive [1,11]. To identify molecular components
involved in plastid development during the youngest
phase of the Arabidopsis life cycle, a screen was conducted
to identify mutants that specifically influence plastid
development in embryos and seedling tissues derived
from the embryo, but not in tissues derived from the api-
cal meristem. Mutants were identified that exhibited coty-
ledon and hypocotyl albinism upon germination due to
improper chloroplast development, while photosynthetic
tissues derived from the shoot apical meristem were green
and appeared to develop normal chloroplasts. This paper
describes mutants in a gene that encodes for a plastid-
localized elongation factor G (EF-G). One mutant allele
from an EMS-(ethyl methanesulfonate) mutagenized
population was found to be the result of the same nucle-
otide substitution responsible for the recently described
snowy cotyledon 1 (sco1) [GenBank:NM_104952] mutant
[12]. Two alleles with T-DNA insertions directly in this
gene resulted in embryo lethality, demonstrating that this
EF-G is essential during embryogenesis. Our analysis of
the different mutant alleles of SCO1 indicate that differen-
tiation of eoplasts to chloroplasts during germination
may have different requirements for protein translation
than for proplastid to chloroplast differentiation in cells
derived from the apical meristem.
Results
Embryo-lethal sco1 alleles
Two alleles of sco1 (sco1-2 and sco1-3) isolated from the
Salk T-DNA insert collection have T-DNA inserts at the C-
terminal end of the gene (Figure 1). In sco1-2, all of the T4
seed tested developed green cotyledons. Genotypic analy-
sis of the viable progeny showed that these plants con-
sisted of a segregating population with a ratio of 1:2 for
wild-type to heterozygote for the T-DNA insert in the
SCO1 gene. No viable progeny were found that were
homozygous for the T-DNA insert. Moreover, examina-
tion of developing siliques on plants that were hetero-
zygous for the sco1-2 T-DNA insert revealed that
approximately 25% of the developing ovules were white
(88 white ovules out of 360 examined; see Figure 2B). The
9 d-old white ovules were similar in size to the green
ovules, but upon dissection they did not appear to contain
a developing embryo. This was presumably due to abor-
tion of the embryo at a very early stage of embryogenesis.
Similar results were observed for the sco1-3 T-DNA inser-
tion line (Figure 2C). These findings indicate that null
mutants in SCO1 are lethal early during embryo develop-
ment and that the EF-G encoded by SCO1 is essential for
plant development.
Conditional seedling-lethal sco1 alleles
We identified one EMS-derived sco mutant that, upon
mapping and sequencing, was found to be identical to the
sco1 allele recently identified by Albrecht et al. [12], where
a G to A base change converted a conserved glycine resi-
due to an arginine. We have designated this allele as sco1-
1 (Figure 3B). Seedlings of sco1-1 rarely survived past the
cotyledon stage unless they were provided with supple-
mentary carbon (Table 1). In addition, an allele (sco1-4)
with a Salk T-DNA insert located 14 base-pairs upstream
of the SCO1 ATG start site was found to have a similar
seedling-lethal phenotype. However, sco1-4 plants could
also be rescued when germinated on medium supple-
The SCO1 mutational map
Figure 1
The SCO1 mutational map. SCO1 encodes for a pre-
dicted protein with a high degree of similarity to an EF-G
containing a chloroplast localization signal (At1g62750). The
locations of the 4 mutant alleles are indicated. The EMS-
mutagenized sco1-1 allele represents a G to A base change
within the GTP-binding domain of the gene, which converts a
glycine at amino acid 132 to an arginine. The other alleles
[sco1-2 (Salk_046154), sco1-3 (Salk_039084), and sco1-4
(Salk_025112)] were isolated from the Salk T-DNA insert
collection.

Page 3

BMC Plant Biology 2007, 7:37http://www.biomedcentral.com/1471-2229/7/37
Page 3 of 10
(page number not for citation purposes)
mented with sugar (Table 1). Unlike the white cotyledons
of sco1-1 (Figure 3B), sco1-4 seedlings had very pale green
cotyledons (Figure 2D). When germinated on media with
sucrose, the first true leaves that emerged from sco1-4
Characterization of the sco1-1 mutant
Figure 3
Characterization of the sco1-1 mutant. A-J are from 5-
d-old light-grown seedlings, while K and L are from 4-d-old
dark-grown seedlings. Upon seedling germination in white-
light, the cotyledons of sco1-1 (B) appear colorless compared
to wild type (A), but leaves that emerge from the apical mer-
istem are green like wild-type leaves. Chlorophyll autofluo-
rescence and cotyledon cross sections show that sco1-1
cotyledon cells (D and F) are almost completely devoid of
chloroplasts except in cells associated with the vasculature,
while wild-type (C and E) show a normal complement of
chloroplasts in cotyledon cells. The albinism phenotype of
sco1-1 is not always complete and green cells can be found in
some sco1-1 cotyledons, where they are typically located
along the margins of the tissue (G). A cross section of a coty-
ledon (H) from such a variegated mutant shows cells with a
normal complement of chloroplasts adjacent to cells devoid
of chloroplasts. Ultrastructural analysis of chloroplasts in
these 'sectored' sco1-1 cotyledon mesophyll cells (J) showed
that they are similar to chloroplasts in wild-type cotyledons
(I). Starch deposition in 4 d-old dark-grown wild-type (K) and
sco1-1 hypocotyls (L) appears similar, indicating that amylo-
plast development is not severely affected in the sco1-1
mutant. Scale bars in I and J are 1 μm.
Phenotypes of T-DNA insertion alleles
Figure 2
Phenotypes of T-DNA insertion alleles. In 9-d-old sil-
iques, wild-type Arabidopsis ovules were green (A), but in sil-
iques from sco1-2 (B) and sco1-3 (C) heterozygotes, white
ovules were found intermixed with normal green ovules,
indicating that chloroplast development is disrupted during
embryogenesis in these alleles. White ovules accounted for
approximately 25% of the total observed (sco1-2, 88 white
ovules out of 360; sco1-3, 36 white ovules in 166). The
upstream T-DNA insertion in 5-d-old sco1-4 seedlings (D,
right) resulted in significantly stunted growth and pale cotyle-
dons when compared to wild-type seedlings of similar age
(D, left).

Page 4

BMC Plant Biology 2007, 7:37http://www.biomedcentral.com/1471-2229/7/37
Page 4 of 10
(page number not for citation purposes)
seedlings were initially pale, but when transplanted to
soil, the seedlings were able to survive and the rosette
leaves of adult plants resembled wild type (data not
shown).
Plastid development in the sco1 mutant
Although most sco1-1 cotyledons appear completely
white, fluorescence microscopy revealed some red
autofluorescence is present in all mutant cotyledons, espe-
cially along the vasculature (Figure 3D). The red autoflu-
orescence, which is indicative of the presence of
chlorophyll, suggested that chloroplast development was
not completely blocked in the mutant seedlings despite
the visually albino appearance. Also, upon closer exami-
nation of some sco1-1 seedlings, we often observed
patches of green cells, which were typically located near
the margin of the cotyledons and in the upper hypocotyl
(Figure 3G). The extent of this sectoring phenotype varied
from seedling to seedling. In the absence of supplemental
sugar, over 80% of the sco1-1 seedlings failed to produce
true leaves and died (Table 1). The mutant seedlings that
were able to survive without the supplemental carbon
source typically had larger patches of green cells in their
cotyledons and/or hypocotyls. These patches of green cells
were presumably capable of providing the seedling with
photosynthate and other essential plastid-derived compo-
nents necessary for survival until the first true green leaves
could develop from the apical meristem. Similarly, sco1-1
seedlings that were able to survive when germinated in
soil had large green patches (data not shown). Leaves and
other photosynthetic tissues derived from the apical mer-
istem in sco1-1 were green and visually indistinguishable
from wild-type plants (Figure 3B).
To evaluate the structural development of chloroplasts in
the cotyledons, we examined cells by light and transmis-
sion electron microscopy in 5-d-old wild-type and sco1-1
cotyledons. As expected, cotyledon mesophyll cells of
light-grown wild-type seedlings contained numerous
well-developed chloroplasts (Figure 3E). In contrast, chlo-
roplasts were essentially absent from cotyledon meso-
phyll cells of typical sco1-1 seedlings except in the bundle
sheath cells that surround the cotyledon vasculature (Fig-
ure 3F), consistent with the appearance of red autofluores-
cence (Figure 3D). In thin-sections from 'green-sectored'
sco1-1 cotyledons (Figure 3H), we observed chloroplast-
containing mesophyll cells directly adjacent to cells that
are devoid of chloroplasts. The chloroplasts that devel-
oped in green sco1-1 cells appeared normal and showed
characteristics of typical wild-type chloroplasts (Figure
3J).
Since chloroplast development was altered in sco1-1 seed-
lings, we stained seedlings for starch to determine if amy-
loplast development was also altered. Starch staining
revealed that sco1-1 hypocotyls contained starch grains,
indicative of the presence of amyloplasts in the endoder-
mis (Figures 3K, L). Starch grains in sco1-1 root columella
cells also appeared similar to those in wild type (data not
shown). Consistent with the role of amyloplasts in gravity
perception [13], gravitropism of hypocotyl, root, and
inflorescence in sco1-1 was found to be similar to wild
type (data not shown). These data indicate that amylo-
plast development is normal in the sco1-1 mutant seed-
lings.
Transcript abundance in conditional-lethal sco1 alleles
Given the location of the genetic lesions in sco1-1 and
sco1-4, we wanted to determine if the relative abundance
of SCO1 transcript in each mutant was related to the dif-
ferent albinism phenotypes of their cotyledons. Using
primers specific to SCO1, we determined that SCO1 tran-
script abundance was similar in wild-type and sco1-1 seed-
lings (Figure 4). Since the EMS mutation in sco1-1
converts a glycine contained within the GTP-binding
domain to an arginine, transcription was expected to be
similar to wild-type. However, the level of transcript in
sco1-4 is significantly reduced, which is consistent with
the location of the T-DNA insert in the promoter region of
SCO1. The level of transcript amplification of ubiquitin
(UBQ) was similar in the mutants and wild type. We also
found that SCO1 protein tagged with GFP was targeted to
chloroplasts and to several non-photosynthetic plastids
(data not shown), confirming data presented by Albrecht
et al. [12].
Embryo development in wild type and sco1
Since Arabidopsis embryos are green during much of their
growth, we examined embryos of sco1-1 plants to deter-
mine if chloroplast development was impaired in the
mutant during embryogenesis. Embryos dissected from
the middle of siliques between 5 and 15 days after fertili-
zation (DAF) showed that sco1-1 and wild-type embryos
were similar in both morphology and developmental rate
(representative embryos from days 8 and 14 are shown in
Figure 5). Embryos from both sco1-1 and wild type were
Table 1: Effect of sucrose on survival of wild-type and sco
mutants. On 0.5-strength MS media supplemented with 2%
sucrose, wild-type, sco1-1, and sco1-4 seedlings have comparable
survival rates. In the absence of sucrose, less than 20% of the
sco1-1 seedlings were able to produce true leaves and survive
into adulthood, and no sco1-4 seedlings survived. Typically, the
sco1-1 mutant seedlings that survived to adulthood in the
absence of a supplemental carbon source were ones that had
larger patches of green cells in their cotyledons and hypocotyls
Genotype+ Sucrose- Sucrose
Col-0
sco1-1
sco1-4
95/98 = 97%
84/92 = 91%
25/26 = 96%
84/94 = 89%
17/89 = 19%
0/21 = 0%

Page 5

BMC Plant Biology 2007, 7:37http://www.biomedcentral.com/1471-2229/7/37
Page 5 of 10
(page number not for citation purposes)
observed to become visibly green around 6 DAF, and
remained so until approximately 12 DAF when the chlo-
roplasts began to dedifferentiate in preparation for dehy-
dration and maturation of the seed. sco1-1 embryos
dissected from siliques that were 11,12, or 13 DAF and the
embryos were still green, were able to develop green
hypocotyls and cotyledons when precociously germinated
on agar growth medium, but embryos dissected after the
embryos had turned white (>13 DAF) developed the char-
acteristic white cotyledons seen in the sco1-1 mutant (data
not shown). At 12 DAF, Arabidopsis embryos are in the
early stages of desiccation and the onset of dormancy, and
chloroplasts are beginning to dedifferentiate [11]. Since
the sco1-1 phenotype could be rescued by bypassing the
maturation stage of embryogenesis, the function of the
mutant EF-G appears to be particularly critical during late
stages of embryo development when eoplasts form.
Expression levels of the Arabidopsis EF-Gs
The most current annotation of the Arabidopsis genome
predicts a total of three nuclear-encoded EF-Gs. Unlike the
plastid-targeted SCO1, the two other EF-Gs (At1g45332
[GenBank:NM_103595]
and
Bank:NM_130067]) contain predicted mitochondrial-tar-
geting sequences [14]. According to the subcellular
prediction program TargetP, their targeting sequences
may allow for dual targeting of the proteins to the mito-
chondria and plastids. During the course of Arabidopsis
development, SCO1 is the most highly expressed of the
EF-Gs, with expression levels peaking at 9.2 times the lev-
els of At1g45332 and At2g45030 in cotyledon tissue (Fig-
ure 6; data compiled from Genevestigator [15]).
Transcript levels of SCO1 are reduced in adult rosette tis-
sue as compared to cotyledon tissue, whereas At1g45332
and At2g45030 mean expression levels remain relatively
constant throughout development, but are always much
lower than for SCO1.
At2g45030 [Gen-
Discussion
sco1 encodes for a translation elongation factor G
In general, the chloroplast genome encodes for genes that
can be classified into several functional categories, includ-
ing genes specific to transcription and translation within
the plastid, photosynthetic genes, and genes involved in
the synthesis of metabolic compounds [16]. Many of the
components of the chloroplast proteome, however, are
nuclear-encoded [17], including a number of factors that
have been shown to be important in regulating transla-
tion of plastid genes [18,19]. For example, studies in vari-
ous plant species indicated that protein initiation factors
and elongation factors, including elongation factors EF-G
and EF-Tu, are present in the nuclear genome and contain
chloroplast-targeting sequences [20-25]. Much of our
knowledge of plastid gene function in transcription and
translation has drawn from structural and functional sim-
ilarities to prokaryotic proteins that serve in a similar
capacity. The presence of plastid-specific ribosomal pro-
teins (PSRPs), however, indicates that at least some
aspects of the translation mechanism in chloroplasts is
unique to plants [26,27].
As shown here and by Albrecht et al. [12], the nuclear
SCO1 gene encodes for a protein translation elongation
factor G with a chloroplast-targeting signal sequence.
Although biochemical activity has not been directly dem-
onstrated for SCO1, similar functionality seems highly
likely since the predicted amino acid sequence of SCO1 is
over 50% identical to the E. coli EF-G fusA [Gen-
Development of sco1-1 embryos
Figure 5
Development of sco1-1 embryos. Representative
embryos were dissected from the middle of siliques on days
8 and 14 after anthesis. With respect to developmental rate
and morphology, sco1-1 embryos were similar to wild type.
The sco1-1 embryos were green from days 6 to 12, but
appeared slightly paler compared to wild-type embryos of
similar age. DAF = days old after fertilization.
Expression of SCO1
Figure 4
Expression of SCO1. RTPCR analysis of the EF-G tran-
script from wild-type, sco1-1, and sco1-4 seedlings demon-
strates a lower abundance of transcript level in sco1-4. The
TDNA insert in sco1-4 is located 14 base-pairs upstream of
the ATG start site, directly affecting its transcription rate.
The genetic lesion in sco1-1 does not appear to affect the
level of SCO1 mRNA, as its level appears similar to that of
wild-type. The level of the loading control UBQ transcript is
similar between all three samples.