Mol Gen Genet (1984) 194:15-23
© Springer-Verlag 1984
The D N A of Arabidopsis thaKana
Leslie S. Leutwiler*, Barbara R. Hough-Evans and Elliot M. Meyerowitz
Divison of Biology, California Institute of Technology, Pasadena, California 91125, USA
Summary. Arabidopsis thaliana is a small flowering plant
of the mustard family. It has a four to five week generation
time, can be self- or cross-pollinated and bears as many
as 104 seeds per plant. Many visible and biochemical muta-
tions exist and have been mapped by recombination to one
of the five chromosomes that comprise the haploid karyo-
type. With the experiments reported here we demonstrate
that Arabidopsis has an extraordinarily small haploid ge-
nome size (approximately 7 x 1 0 7 nucleotide pairs) and a
low level of cytosine methylation for an angiosperm. In
addition, it appears to have little repetitive D N A in its
nuclear DNA, in contrast to other higher plants.
Introduction
Plants have large differences in their nuclear D N A content.
Within angiosperms there is a nearly thousand-fold range
of variation, and there appears to be no correlation between
genome size and organismal complexity (Bennett and Smith
1976). Most of the angiosperms currently used in molecular
genetic studies have large genomes; correlating with these
large (more than 109 nucleotide pairs) genomes is a large
fraction of D N A that is repeated many times, with individ-
ual repetitive elements dispersed throughout the genome
(Flavell 1980). Both of these properties, large haploid ge-
nome and high fraction of dispersed repetitive DNA, create
difficulties in the standard procedures of molecular genet-
ics: a large genome requires that large genomic libraries
be screened to obtain any individual gene, and collection
of overlapping clones to define a chromosomal region is
greatly hindered by dispersed repetitive elements.
Arabidopsis thaliana is a small flowering plant of the
mustard family with a short generation time (four to five
weeks), the ability either to cross- or self-pollinate, and a
haploid chromosome number of only five. It has been used
in a number of experiments in classical and biochemical
genetics (R6dei 1970). Numerous visible and biochemical
mutations are known, and many of these have been mapped
by recombination to positions on the five linkage groups
(Koorneef et al. 1983). It is susceptible to infection by Agro-
bacterium tumefaciens; there thus exists a potential for using
* Current Address: Department of Biology, University of Califor-
nia, Los Angeles, CA 90024, USA
Offprint requests to: E.M. Meyerowitz
Ti plasmid as a means of achieving DNA-mediated trans-
formation of Arabidopsis (Aerts et al. 1979). Microspectro-
photometric measurements of nuclear D N A content in this
plant indicate that it has a small haploid genome (Bennett
and Smith 1976). Thus, Arabidopsis has several advantages
as an organism for use in molecular genetic studies of higher
plants. As a preliminary to a number of such studies, we
here report experiments that characterize the composition
and organization of Arabidopsis DNA, including measure-
ments of kinetic complexity, genome size, and nucleotide
composition. We find that Arabidopsis has an extraordinari-
ly small genome (approximately 7 x 107 nucleotide pairs)
and little repetitive DNA. This confirms its potential as
an organism for use in molecular genetic experimentation,
and provides basic information as a foundation for such
studies. In addition, it establishes a new minimum genome
size for a higher plant.
Materials and methods
Biological material. Seeds of Arabidopsis thaliana strain Col-
umbia were obtained from Dr. A. Kleinhofs, Program in
Genetics, Washington State University, Pullman, WA
99164. Wheat germ (Fisher Natural - not toasted) was ob-
tained from a local market.
DNA isolation. Arabidopsis seeds were germinated on a mix-
ture of sterile soil and peat (mixed 3:1) and grown under
constant illumination (7000 lux) at 25 ° C and 60% relative
humidity. Approximately 10 g of five week old plants were
harvested, rinsed with water, and ground in a mortar and
pestle with an equal weight of glass beads in the presence
of liquid nitrogen. Two grams of the powder were added
to 10 ml 150 mM Tris-HC1, pH 8.5; 100 m M EDTA; 2%
N-lauroyl sarkosine, sodium salt; 0.1 mg/ml proteinase K
and incubated at 37 ° C with gentle stirring for 30 min. The
residue was removed by centrifugation and re-extracted
twice; the supernatant was ethanol precipitated to remove
a fluorescent compound which interfered with the ethidium
bromide fluorescence of subsequent steps. Following cen-
trifugation of the ethanol precipitate, the pellets were resus-
pended in 10 mM Tris-HC1, pH 8.0; 1 mM E D T A (TE)
and banded to equilibrium in CsC1 density gradients in
the presence of 1 mg/ml ethidium bromide. The bands were
collected, the ethidium bromide was removed by butanol
extraction, and the D N A ethanol precipitated. The pellets
were resuspended in 1 ml TE, loaded on a i M NaC1 solu-
16
tion and centrifuged in a SW 50.1 rotor at 40,000 rpm for
5 h in order to remove any residual RNA. The D N A pellet
was resuspended in TE, and the D N A concentration was
measured by absorption at 260 nm minus that at 320 nm
to correct for light scatter.
Wheat germ D N A was prepared essentially according
to the above procedure. Wheat germ (0.5 g) was ground
in an ice cold mortar and pestle with an equal weight of
glass beads in the presence of 3 ml of proteinase K (10 rag)
in 150 m M Tris-HC1, pH 8.5, 100 m M EDTA. Following
grinding, the slurry was made up to 10 ml with N-lauroyl
sarkosine (to a final concentration of 2%) in the Tris-
E D T A mixture and incubated at 37 ° C with gentle stirring
for 30 min. The remaining procedure for D N A isolation
was the same as that for Arabidopsis DNA, except the initial
ethanol precipitation step was omitted.
Chloroplast DNA isolation. Whole plants (8.75 g) were dark
adapted 12-18 h and homogenized in 40 ml buffer A (0.3 M
mannitol, 0.05 M Tris, 0.003 M EDTA, 0.001 M mercapto-
ethanol, 0.1% bovine serum albumin, pH 8.0 - Kolodner
and Tewari 1975) with three 5-second bursts at high speed
in a Waring blender. Following filtration through 4 layers
of Miracloth, the supernatant was layered on 2 M su-
crose: 80% Percoll: 60% Percoll:40% Percoll in buffer A
(0.6:1:1:1 v/v) and centrifuged at 8,000 rpm in a Sorvall
HB-4 rotor for 30 rain. The chloroplast band at the
40%-60% Percoll interface was collected, diluted in buffer
A and pelleted in a HB-4 rotor at 2,000 rpm for 5 min.
The pellet was solubilized in 5 ml 0.05 M Tris-HC1, pH
8.0, 0.02 M E D T A containing 2% N-lauroyl sarkosine and
100 gg/ml proteinase K, incubated at 37 ° C 20 min and phe-
nol extracted twice. The aqueous phase was re-extracted
with chloroform: 1% isoamyl alcohol and ethanol precipi-
tated. The D N A was further purified on a CsC1 equilibrium
density gradient.
Preparation of labeled probes. A. thaliana total D N A or
Drosophila melanogaster embryo D N A was nick translated
(Rigby et al. 1977) with [3H]-deoxycytidine triphosphate,
and a lambda clone containing a chloroplast D N A insert
(2bAt003) was nick translated with [32p]-deoxycytidine tri-
phosphate. The nick translation mixture was incubated at
13°C for 45 min and the D N A purified by passage over
a Bio-Gel P-60 (100-200 mesh; Bio-Rad) column.
The nick translation yields a mixture of fragments rang-
ing in size from greater than 1700 base pairs (bp) down
to less than 154 bp. Therefore, it was necessary for the reas-
sociation measurements to isolate single-stranded D N A
fragments longer than 500 bp so that they (1) could be
sheared uniformly with the unlabeled D N A and (2) were
long enough to reassociate under the hybridization condi-
tions. D N A from the plasmid pBR325 (Bolivar 1978) was
digested with Hint and used as a standard for size measure-
ments. In order to make the nick translated D N A single-
stranded the D N A probe was made 0.1 N in N a O H prior
to loading. Following electrophoresis through 2% agarose
gels to separate the different size D N A strands, a slot was
cut in the gel perpendicular to the electric field in the track
containing the nick translated D N A and at the location
for fragments of 500 bp. A piece of Whatman DE 81 paper
was inserted and electrophoresis continued until all activity
present initially in the gel on fragments greater than 500 bp
could be found on the DE 81 paper. The piece of DE 81
paper was then rinsed three times with 0.1 M Tris-HC1,
pH 7.5, 0.1 M NaC1 and the D N A eluted with 0.1 N NaOH.
Shearing of DNA. Unlabeled Arabidopsis (and Drosophila)
D N A and labeled tracer D N A of the correct size (described
above) were mixed before shearing in a Virtis homogenizer
as described by Britten et al. (1974). Ten ml of the D N A
mixture in TE was diluted with 20 ml glycerol and sheared
in a Virtis homogenizer at 60,000 rpm in a dry ice-ethanol
bath for 30 rain. The average single-strand length of the
D N A was determined by electrophoresis through a 2%
agarose gel using Hinf I digested pBR325 D N A as size
standards. The unlabeled D N A was visible by UV fluores-
cence while the [32p]-labeled D N A was visualized by auto-
radiography of the gel. Both procedures indicated the D N A
had been sheared to an average single-strand length of 375
nucleotides. After shearing the D N A was ethanol precipi-
tated and the precipitate subsequently dissolved in a small
volume of TE. This solution was layered on a Chelex-100
(200-400 mesh, sodium form; Bio-Rad) column in order
to remove heavy metal ions; the fractions containing D N A
were combined and reprecipitated with ethanol. The precip-
itate was resuspended in 0.4 M sodium phosphate buffer
pH 6.8 (PB) and the D N A concentration determined.
DNA reassociation. Reaction mixtures were prepared and
analyzed essentially according to Britten et al. (1974). D N A
samples ranging in concentration from 3.33 gg/ml to
1,375 gg/ml in either 0.12 M or 0.4 M PB were sealed in
capillaries or ampoules, denatured by boiling for 1 min,
and reassociated in a 60°C water bath for the required
incubation times. After reassociation to the appropriate
Cot, the samples were frozen in dry ice-acetone and kept
frozen at - 2 0 ° C.
Reassociated D N A was separated from single-stranded
D N A by hydroxylapatite (Bio-Rad, D N A Grade Bio-Gel
HTP) chromatography in water-jacketed columns. D N A
samples reassociated in 0.12 M PB were thawed and loaded
directly on the column (2 ml bed volume), while those sam-
ples reassociated in 0.4 M PB were diluted to 0.12 M PB
before loading. Single-stranded D N A was eluted in five to
six 1 ml fractions with 0.12 M PB at 60 ° C; double-stranded
D N A bound to hydroxylapatite was denatured by heating
the column to 98 ° C and eluted in five 1 ml fractions with
the same buffer. Reassociation of unlabeled D N A was de-
termined by measuring the A26 o nm for each fraction and
correcting for light scattering at 320 nm. Reassociation of
labeled D N A was determined subsequently by counting
each fraction in Aquasol-2 (New England Nuclear) in a
Beckman LS-250 scintillation counter. Tritium and [32p]_
cpm were measured simultaneously by counting in a half-
tritium channel and a channel spanning the [14C]- [32p]
energy range respectively.
Cot values were calculated for each sample by multiply-
ing the D N A concentration (moles nucleotide per liter) by
time (sec). Reassociation data were fit by the equation:
r
C~
1
Co 1 + kCo t
where C = concentration of single-stranded D N A at time
t, Co = original D N A concentration, and k = second order
reassociation rate constant. The results were analyzed by
a non-linear least-squares computer program (Pearson et al.

1977) designed to fit theoretical second order components
to the observed reassociation kinetics.
Thermal denaturation of DNA. An aliquot of D N A from
experiment A (Table 1) was reserved prior to shearing in
order to determine the GC content of the D N A preparation.
A. thaliana D N A at a concentration of 12.5 gg/ml and a
control sample of calf thymus D N A at a concentration
of 42.7 lag/ml both in 0.12 M PB were melted in a Beckman
Acta double-beam spectrophotometer equipped with a jack-
eted cuvette holder and automatic sample changer. A refer-
ence cuvette contained 0.12 M PB in order to ensure a linear
base line.
The temperature was raised to 45 ° C and then increased
at the rate of 1.5°/rain to 95 ° C. Separation of strands was
monitored by the increase in absorbance at 260 nm. The
hyperchromicity was calculated as the percent of final ab-
sorbance; the melting temperature (Tin) was that tempera-
ture which produced a 50% increase in hyperchromicity.
The mol percent G + C content of native D N A was calcu-
lated using the equation of Felsenfeld (1971):
% G + C=2.44 (Tm- 69.35)
Determination of percent of 5-methylcytosine. The percent
of 5-methylcytosine in 130 lag A. thaliana, 200 gg D. melano-
gaster and 160 lag wheat germ D N A was determined by
lyophilizing the DNAs separately and then hydrolyzing to
bases by heating at 180° C for 25 rain in sealed glass am-
poules containing 200 lal formic acid. The hydrolysates were
evaporated to dryness under nitrogen and resuspended in
0.1 M HC1 in a ratio of 1 lag starting DNA: 1 gl 0.1 M
HC1.
The bases released by formic acid hydrolysis were ana-
lyzed by HPLC (Diala and Hoffman 1982; Diala et al.
1981). Samples (25-40 lal) were injected into an Altex Ultra-
sil-10 CX column and eluted at ambient temperature with
0.02 M ammonium phosphate (monobasic) buffer, pH 2.3.
Cytosine and methyl cytosine bases were identified relative
to the elution of authentic standards with detection at
280 nm, and the data were processed by a Shimadzu C-R1A
chromatopac integrator and calculated using standard
curves. Each value reported is the mean of three indepen-
dent measurements.
Results
G + C content ofArabidopsis DNA
Arabidopsis thaliana D N A was isolated from whole plants
by the procedures detailed in Materials and Methods. Since
this D N A was to be used in all subsequent experiments,
its purity was checked by observing its denaturation in a
temperature-controlled spectrophotometer cell. The D N A
melted in 0.12 M PB with Tm=86.3 ° C and no hyperchro-
micity was observed at temperatures lower than the D N A
denaturation transition, indicating the absence of any R N A
or single-stranded D N A in the preparation. The hyperchro-
micity on melting, calculated as percent of final (single-
strand) absorbance, was 22%. The G + C content of D N A
can be calculated from its T m (Felsenfeld 1971); the value
obtained for Arabidopsis D N A is 41.4% G + C, well within
the range (35.6% to 49.1% G + C) known for angiosperms.
The Arabidopsis figure of 41.4% is almost identical to the
41.2% G + C content of Brassiea oleifera (Vanyushin and
17
Belozerskii 1959), and close to the 45% G + C content of
Brassica oleracea (Thomas and Sherratt 1956), both
members of the same family as Arabidopsis thaliana. As
a control for this denaturation experiment, calf thymus
D N A was melted in parallel with the plant DNA. It melted
at T m= 86.7 ° C. This T m indicates a G + C content of 42.3%,
comparable to published values for calf thymus D N A (40.0
to 45.0%, Shapiro 1976). Since the Arabidopsis D N A was
found to be pure and of average G + C content, D N A reas-
sociation cound be carried out under standard conditions
(Britten et al. 1974).
Reassociation kinetics of Arabidopsis DNA
The first experiment to characterize the sequence organiza-
tion and kinetic complexity of total Arabidopsis thaliana
D N A was to measure the reassociation rate of denatured
D N A under defined conditions. Plant D N A was sheared
to an average 375 nucleotides long, denatured, and allowed
to reassociate. The percent reassociation was measured at
a number of time points by separating single- and double-
stranded D N A on temperature-controlled hydroxylapatite
columns and then quantitating the amount of each compo-
nent by UV light absorption at 260 nm. The reassociation
values are shown in Fig. 1. The curve drawn through the
measured points in Fig. 1 is a computer best fit to second
order reassociation kinetics (see Materials and Methods).
The best fit consisted of two kinetic components. In addi-
tion, a component of the D N A was reassociated by the
time the first measurements were made (Cot = 0.002). Thus,
three components were discovered: a rapidly reassociating
component, which comprises 10% of the plant DNA, a
middle repetitive component containing 27% of the DNA,
and a slowly reassociating component with 55% of the
DNA. The remaining 8% of the D N A did not reassociate;
this D N A presumably consisted of small sheared fragments
unable to form stable duplexes which will bind to hydroxy-
lapatite. Table 1, Part A, displays the results of the analysis
of this curve, including a remarkably small value for the
overall kinetic complexity of the plant DNA.
In addition to the unlabeled Arabidopsis D N A present
in this reassociation experiment, small quantifies of [3H]-
labeled Arabidopsis DNA, sheared to the same size distribu-
tion as the bulk unlabeled DNA, were included. The reasso-
ciation of this tracer, driven by the unlabeled DNA, was
measured as was the reassociation of the bulk DNA, except
that quantitation after the hydroxylapatite column was by
scintillation spectrometry. This labeled tracer served as a
control to be certain that the results derived from optical
density measurements did not reflect unknown UV absorb-
ing substances contaminating the DNA. They did not. As
demonstrated in Table 1, Part A, the [3H] tracer results
parallel those obtained from the unlabeled DNA.
The D N A used in this experiment was isolated from
whole plants, and not only from nuclei. Consequently, chlo-
roplast D N A would be expected to make up a significant
fraction of the total D N A : in Sinapis alba, a member of
the mustard family related to Arabidopsis, the chloroplast
genome is 1.58 × 105 bp; the complexity of this genome is
1.34 × 105 bp, since about 2.4 × 104 bp, including the ribo-
somal RNA coding regions, are represented twice (Link
et al. 1981). If the size of the Arabidopsis chloroplast ge-
home is similar, and there are 20 to 80 chloroplasts per
green cell (R6dei 1973), and 20 to 60 genomes per chloro-
18

0
<~ O
Z
D 20
~6 o o o
g
~- 4 0
O
"O
O
6O

§ so
o_ I00
-3 -2 -I 0 I 2 3 4
Log equivolent Col
Fig. 1. Reassociation kinetics of total Arabidopsis D N A fragments. 400 gg unlabeled plant D N A was combined with 0.14 gg labeled
size selected whole plant D N A (specific activity: 5.6x 105 cpm [aH]/gg) and 0.14 ng labeled sized 2bAt003 D N A (specific activity:
1.7 x 108 cpm [32p]/pg)). The D N A was sheared to an average single-strand length of 375 nucleotides, resuspended in 0.4 M ( C o t = > 50)
or 0.12 M ( C o t = <50) sodium phosphate, pH 6.8 (PB) and denatured by boiling for 1 rain. Samples were reassociated at 60°C to
the appropriate Cot value and frozen in dry ice-acetone. All samples were made up to 1 ml in 0.12 M PB immediately prior to loading
(except the Cot = 0.02 and 0.002 samples which were already in 3 ml 0.12 M PB). The reassociated D N A was fractionated by hydroxylapa-
tite (HAP) chromatography and the relative amounts of double- and single-stranded plant D N A were determined by absorbance at
260 nm minus 320 nm. The curve drawn through the points represents a least squares fit for two components, assuming second order
kinetics and allowing all the parameters to free float (rms=0.021). The lower curves (dashed lines) represent the predicted reassociation
kinetics for the pure middle repetitive and single copy components. The reassociation of labeled DNAs is not shown in the figure.
The kinetic parameters for the reassociation of the unlabeled Arabidopsis DNA, [3H]-labeled Arabidopsis D N A and [32p]-labeled 2bAt003
D N A are presented in Table I A

Table 1. Kinetic analysis of Arabidopsis D N A reassociation

D N A prepn. Component Fraction Reassociation Complexity Root mean

of D N A rate constants (bp) square error
in whole D N A
(M -1 s -1)

A Arabidopsis( A 2 6 o ) Foldback and highly repet. 0.10

Middle repet. 0.27 4.64 6.0 x 104 0.021
Single copy 0.55 0.0080 7.0 x 107
Arabidopsis [3H] Foldback and highly repet. 0.14
Middle repet. 0.23 4.78 4.9 x 104 0.009
Single copy 0.50 0.012 4.2 x 10 v
2bAt003 [32p] 1.93 1.25 x 105 0.005

B Drosophila [3HI Foldback and highly repet. (0.26)

Middle repet. 0.21 0.44 4.84 x 105 0.017
Single copy 0.45 0.0042 1.10 x 108
2bAt003 [32p] 1.82 0.018

C Arabidopsis[3H] Foldback and highly repet. (0.23)

Middle repet. 0.26 3.29 8.1 x 104 0.012
Single copy 0.47 0.0089 5.4 x 107
XbAt003 [32p] 2.44 1.2 x 105 0.013

The separation of kinetic components, and calculation of the rate constant and complexity of each component was as described in
Materials and Methods, and in Britten et al. (1974). The highly repetitive fractions in the reassociation curves described in B and
C are in parentheses to indicate that these measurements are overestimates: in the experiment detailed in part A the samples that
provided the initial reassociation points were diluted to 3 ml before the start of the experiment to avoid reannealing during the loading
of the hydroxylapatite column. The experiments of parts B and C were not diluted sufficiently, since the information desired was
that from reassociation of single-copy sequences. The complexity of the chloroplast D N A hybridized by 2bAt003 in preparations A
and C was calculated using the reassociation rate constant of the cloned tracer sequences and the genome fraction of the middle
repetitive component of the Arabidopsis D N A reassociation curve

plast ( B e d b r o o k a n d K o l o d n e r 1979), t h e n the a m o u n t o f D N A , is i n d e e d largely or totally due to this n o n - n u c l e a r

c h l o r o p l a s t D N A p e r green cell w o u l d be e x p e c t e d to be
f r o m 6 x 107 to 8 x 10 s bp. It was t h e r e f o r e i m p o r t a n t to
d e t e r m i n e if the m i d d l e - r e p e t i t i v e c o m p o n e n t seen in the
r e a s s o c i a t i o n curves, w h o s e c o m p l e x i t y a n d a m o u n t is
a b o u t e q u a l to t h a t e x p e c t e d o f Arabidopsis c h l o r o p l a s t
set o f sequences. T o d o this, the r e a s s o c i a t i o n e x p e r i m e n t
described a b o v e c o n t a i n e d , in a d d i t i o n to u n l a b e l e d a n d
[3H]-labeled Arabidopsis D N A , a trace q u a n t i t y o f [32p]_
labeled Arabidopsis c h l o r o p l a s t D N A derived f r o m a bacte-
r i o p h a g e 2 r e c o m b i n a n t clone.
 19

This 2 clone, 2bAt003, was isolated from an Arabidopsis

library (Pruitt and Meyerowitz, unpublished) and shown
to contain a portion (about 104 bp) of the chloroplast ge-
nome by the following series of experiments. First, the 2
clone was [32p]-labeled by nick translation, and hybridized
Kbp to a nitrocellulose gel blot filter (Southern 1975) prepared
from Arabidopsis D N A digested with the restriction endo-
nuclease Eco RI. After hybridization and autoradiography,
5.15_ it could be seen that the clone D N A hybridized to plant
4.27 -
D N A fragments identical in size to the Eco RI restriction
fragments o f the cloned D N A , and with an intensity of
:3.5:3 -
hybridization much greater than that typically seen with
single-copy ,~ clones. Thus, the clone contained unrear-
2.0:3 - ranged, repetitive Arabidopsis D N A . To show that this re-
I .58 - petitive D N A was chloroplast D N A , Arabidopsis chloro-
I .:38 - plasts were purified and chloroplast D N A prepared as de-
scribed in Materials and Methods. This D N A was digested
0.8:3 - with EcoRI, subjected to gel electrophoresis, blotted to ni-
trocellulose and hybridized with [32p]-labeled 2bAt003. Au-
toradiography of the blot showed that the clone does hybri-
dize to purified chloroplast D N A (Fig. 2), and that the pat-
tern of hybridization is identical to that seen in the blots
using whole plant D N A .
ab The reassociation of the 2bAt003 insert sequences in
the annealing experiment of Fig. 1 was driven by chloro-
Fig. 2a, b. Identification of 2bAt003 as a chloroplast genomic plast sequences in the unlabeled D N A . The rate of reasso-
clone, a Chloroplast D N A was prepared as described in Materials ciation of the chloroplast sequences in 2bAt003 is similar
and Methods. Approximately t btg of chloroplast D N A was di- to that of the middle repetitive component extracted from
gested for 2 h with 2 units of EeoRI in the presence of 4 m M the plant D N A Cot curves. Thus, the middle repetitive
spermidine and buffer recommended by the supplier (New England D N A likely contains chloroplast sequences. That it is large-
Biolabs). Restricted DNA was separated by electrophoresis on a
ly comprised of chloroplast D N A is indicated by the rela-
0.5% agarose gel, stained with ethidium bromide and photo-
graphed under 302 nm illumination, b The gel was blotted to nitro- tion of the complexity of this component with that expected
cellulose, probed with ZbAt003 DNA (specific activity=2.1 x of chloroplast D N A (Table 1). It should be noted that mito-
108 cpm [3zp]/gg), and exposed for 2 h without intensifier. Molecu- chondrial D N A would not be expected to contribute detec-
lar length standards derived from EcoRI and HindIII digested 2 tably to our whole plant D N A reassociation curve: in sever-
phage DNA are indicated to the left of lane a al species of the genus Brassica, members of the mustard

z 20
~6
4O
Q
6O

,o0 , , , , 1 , ....
-5 -2 -I 0 I 2 3 4
Log equivolenl Col
Fig. 3. Reassociation kinetics of labeled Drosophila and 2bAt003 D N A fragments. Equal amounts (200 lag) of Drosophila and Arabidopsis
unlabeled D N A were combined with 0.1 lag labeled sized Drosophila D N A (specific activity: 8.6 x 105 cpm [3H]/lag) and a.0 ng labeled
sized 2bAt003 D N A (specific activity: 2.4x 107 cpm [azp]/lag) and treated as described in Fig. 1. The relative amounts of double-
and single-stranded tracer D N A were determined by measuring [3H] cpm (Drosophila (e)] and [a2p] cpm (2bAt003 (A)] bound to
HAP at each Cot value. The Drosophila curve represents a least squares fit for two components and allows all parameters to free
float (rms=0.017). The lower curve (dashed line) represents the predicted reassociation kinetics for the pure single copy component.
The 2bAt003 curve represents a least squares fit for one component, allowing all parameters to free float (rms=0.018). The partial
(42%) reannealing of 2bAt003 is the annealing of the chloroplast insert in the clone driven by the plant DNA, the 58% not annealed
is 2 phage DNA, which is present at too low a concentration to reassociate
20

Z
r'q 20 --o-

4o
.{:2_
--
6O
o

g 80

Y_
I00 I I I . . . . . . 4--. I I - -
'- -2 -I 0 I 2 3
Log equivQIent Cot
Fig. 4. Reassociation kinetics of labeled Arabidops# and 2bAt003 DNA fragments. Equal amounts (200 gg) of Drosophila and Arabidopsis
unlabeled DNA were combined with 0.1 gg labeled sized whole plant DNA (specific activity; 6.2 x 105 cpm [3H]/gg) and 4,6 ng labeled
sized 2bAt003 DNA (specific activity: 3.3 x 1 0 6 cpm [32p]/p,g) and treated as described in Fig. 1. The relative amounts of double-
and single-stranded tracer DNA were determined by measuring [3H] cpm [Arabidopsis (o)] and [ 3 2 p ] cpm [2bAt003 (A)] bound to
HAP at each Cot value. The 2bAt003 curve represents a least squares fit for one component, allowing all parameters to free float
(0.013). The Arabidopsis curve represents a least squares fit for two components with the reassociation rate for the middle repetitive
component fixed to the value obtained for the reassociation of 2bAt003 DNA (rms=0.017) The lower curves (dashed lines) represent
the two components

family, mitochondrial D N A has been shown to comprise
only 1%-2% of the total D N A of mature leaves (Vedel
and Mathieu 1982).
To be certain that the low complexity determined for
the slowly reassociating component of Arabidopsis D N A
is correct, and not the result of some contaminant in our
Arabidopsis D N A preparation that causes denatured D N A
to reanneal more rapidly than expected, a set of control
experiments was performed. These controls comprised two
additional reassociation curves, in which the kinetic com-
plexity of Arabidopsis D N A was directly compared to that
of D N A from embryonic nuclei of Drosophila meIanogaster.
Drosophila D N A was chosen for this comparison because
its complexity is known from a large number of different
experiments (Laird and McCarthy 1969; Laird 1971; Scha-
chat and Hogness 1973; Manning et al. 1975; Crain et al.
1976), and because its complexity (1 to 2 x 108 nucleotide
pairs) is near that derived for Arabidopsis DNA, thus pro-
viding a control that gives reassociation points over the
same range of Cot values as does Arabidopsis DNA. Both
control curves were generated with identical mixtures of
equal amounts of unlabeled Arabidopsis and Drosophila
DNA, and trace amounts of [3zP]-labeled 2bAt003 DNA.
The first incubation mixture also contained [3H]-labeled
Drosophila D N A tracer, the second, [3H]-labeled Arabidop-
sis tracer. All of the DNAs were sheared to an average
375 nucleotide single-strand length. The kinetic complexity
of Drosophila and Arabidopsis DNAs were thus measured
in parallel experiments; and in addition the experiment with
Arabidopsis tracer provided a repeat of the original analysis
of this DNA. As shown in Fig. 3 and in Part B of Table 1,
there was no component in the Arabidopsis D N A prepara-
tion that accelerated reassociation: the Drosophila D N A
annealed at the expected rate in the presence of plant DNA.
In addition, as seen in Fig. 4 and in Part C of Table 1,
the repeat Arabidopsis curve showed reannealing of the mid-
dle repetitive and slow components of Arabidopsis D N A
almost identical to that determined in the original experi-
ment. The similar reassociation rates of the [32p] 2bAt003
D N A in the parallel control curves shows that the D N A
in two control experiments annealed the same way, with
no discrepancies caused by experimental artifact (Table 1,
parts B and C).
Methylcytosine in Arabidopsis DNA
Plant D N A usually has a high content of 5-methylcytosine,
with values reported from 2.0% of total nucleotides (9.7%
of cytosine residues) as 5-methylcytosine in Brassica oleifera
(Vanyushin and Belozerskii 1959) to 10% of total nucleo-
tides (33% of all cytosines) as 5-methylcytosine in rye, Se-
tale cereale (Thomas and Sherratt 1956). Since the genomic
complexity of Arabidopsis D N A is unusually low, it seemed
possible that Arabidopsis D N A might also be unusual in
its content of 5-methylcytosine. To find out, total plant
Arabidopsis D N A was digested with the restriction endonu-
cleases HpaII and MspI in parallel preparations, and the
digested D N A subjected to electrophoresis in parallel lanes
of an agarose gel. Both enzymes recognize and cleave the
site C-C-G-G in double-stranded DNA, but MspI cleaves
whether or not the internal cytosine residue is methylated,
while HpaII only cleaves if this base is unmethylated. Fig-
ure 5 shows the patterns of digested and undigested Arabi-
dopsis D N A run in the same agarose gel. No difference
can be seen between the lanes containing HpaII and MspI
digested D N A ; D N A in both of these lanes (whether the
discrete bands seen, which are chloroplast derived, or the
background continuum of chromosomal DNA) has been
cleaved to a much smaller mean size than that of the undi-
gested DNA. This indicates a low level of 5-methylcytosine
at C-G sequences. As a control, and for contrast, wheat
germ D N A was treated as was the Arabidopsis DNA.
Wheat germ D N A has 82% of the cytosines in the sequence
C-G methylated (Gruenbaum et al. 1981), and has a level
of cytosine methylation typical of most angiosperms (Sha-
piro 1976). Figure 5 shows the result: undigested and Hpa
II-treated wheat germ D N A give similar gel patterns, while
MspI-cleaved wheat germ D N A shows a broad distribution
of fragments smaller than those in other lanes.

Kbp
5.53
2.05 - The mole percent 5-methylcytosineis expressed as percent of cyto-
I .58 -
I ,:58 -
0.95 -
0.8:5 -
0.56 - The most striking result obtained in this work is the low
abcdef
Fig. 5. Detection of plant DNA methylation by restriction enzyme
analysis. DNA was prepared as described in Materials and Meth-
ods and digested for t.75 h with enzymes in buffers recommended
by the vendor (New England Biolabs). t gg of DNA and either
2 u of HpaII or 7 u of MspI were used in each digestion. Restricted
DNA was separated by electrophoresis on a 1.5% agarose gel,
stained with ethidium bromide, and the ethidium fluorescence pho-
tographed in 302 nm illumination. Undigested Arabidopsis DNA
is shown in lane a. The adjacent lanes show Arabidopsis DNA
digested with HpaII (lane b) and MspI (lane c). Lane d is undigested
wheat germ DNA, lane e is wheat germ DNA digested with HpaII,
and lane f wheat germ DNA restricted with MspI
Quantitation of 5-methylcytosine levels in Arabidopsis
D N A was achieved by hydrolyzing this D N A to individual
bases, separating the bases by high pressure liquid chroma-
tography, and quantitating levels of cytosine and 5-methyl-
cytosine (Diala et al. 1981; Diala and Hoffman 1982). As
controls Drosophila embryo DNA, which is unmethylated
(Urieli-Shoval et al. 1982), and wheat germ DNA, with 5-
methylcytosine levels reported as being from 23.5% to
25.2% of total cytosines (Shapiro 1976), were processed
as well. The results are shown in Table 2. The Drosophila
and Triticum values obtained are close to the published
values, and Arabidopsis thaliana showed only 4.6% of its
cytosine residues as methylcytosine. Since chloroplast D N A
of plants is generally not methylated (Kung 1977), and chlo-
roplast D N A comprises up to 27% of the total Arabidopsis
D N A used for measurement, the methylcytosine content
21
Table 2
DNA mol %
5-methylcytosine
Arabidopsis thaliana 4.6
Drosophila melanogaster < 0.2
Triticum (wheat germ) 20.1
sine residues that are 5-methylcytosine
of Arabidopsis nuclear D N A may be as high as 6.3% of
all cytosines. This is the lowest value reported for any an-
giosperm.
Discussion
complexity of the slowly reassociating (single-copy) fraction
of the Arabidopsis thaIiana genome, with three different
measurements giving a range of complexities from
4.2 x 1 0 7 bp to 7 x 1 0 7 bp for this fraction. It is possible
to calculate both the total genome complexity and the total
genome size from the reassociation results. Total complex-
ity is essentially equal to complexity of the single-copy frac-
tion, since the complexity of the other fractions is orders
of magnitude lower. Total genome size can be derived from
the single-copy complexity and fraction of the genomic
D N A found in the single-copy component. If we assume
that the middle repetitive component in our reassociation
curves is entirely or almost entirely chloroplast D N A (see
Results), then the value for nuclear genome size derives
only from the highly repeated and nonrepeated D N A frac-
tions. Using the figures from the experiment which mea-
sured reassociation by ultraviolet absorption (Table 1), we
calculate a haploid nuclear genome size of 8.3 x 1 0 7 bp. Us-
ing the results from the [3Hi-labeled tracer in the same
experiment, we get the similar value 5.4 x 1 0 7 bp. The mean
of these measurements gives an estimate of the size of the
Arabidopsis nuclear genome of 7 x 107 bp. Even if the puta-
tive chloroplast D N A is included in the calculations, the
figure arrived at for total cellular genome size is only 40%
larger than the nuclear genome size calculated above: so
that under any set of assumptions, the genome of Arabidop-
sis thaliana is smaller by far than that calculated for any
other angiosperm (Bennett and Smith 1976).
If all of the middle repetitive D N A is indeed chloroplast,
then the number of chloroplast genomes per haploid ge-
home in an average Arabidopsis cell is in the range 350-600,
as calculated from our reassociation experiments; a diploid
cell thus has 700-1200 chloroplast genomes. This value is
consistent with the amount of chloroplast D N A per cell
in Arabidopsis and related plants (see above, and Vedel
and Mathieu 1982). It is also consistent with the relative
intensity of autoradiographic signals from Arabidopsis total
D N A sequences on genome blots that have been probed
with radioactively labeled cloned chloroplast and single-
copy D N A fragments, and with the relative intensity of
the autoradiographic signals from equal amounts of filter-
bound cloned chloroplast and single-copy sequences probed
with radioactively labeled Arabidopsis whole plant D N A
(Pruitt and Meyerowitz, unpublished). While this evidence
indicates that all of the middle repetitive component of the
22
Arabidopsis D N A might be chloroplast D N A , it does n o t
rule out the possibility that other, nuclear middle repetitive
sequences m a y m a k e up a p a r t o f this component. Experi-
ments using cloned Arabidopsis genomic fragments to di-
rectly examine this middle repetitive c o m p o n e n t are now
in progress (Pruitt and Meyerowitz, unpublished). It should
be noted that the values for chloroplast D N A complexity
obtained from total D N A are lower by a factor o f two
than the value obtained from molecular studies o f the chlo-
roplasts o f related plants, and t h a t the second order reasso-
ciation rate constant measured for reassociation o f the mid-
dle repetitive Arabidopsis c o m p o n e n t is higher than that
directly measured for Arabidopsis chloroplast D N A using
the [32p]-labeled chloroplast clone. I f the rate constant ob-
tained for the middle repetitive Arabidopsis c o m p o n e n t is
fixed to a value that is equal to that shown b y the chloro-
plast tracer a n d thus to a value that gives a m o r e typical
chloroplast complexity, instead o f leaving all parameters
free in our c o m p u t e r analysis, a n d then the best second-
order reannealing curves fit to the measured points, we
get the following result from the [3H]-traced reassociation
experiment shown in P a r t A o f Table 1 : foldback a n d highly
repetitive D N A are still 14% o f total D N A , middle repeti-
tive D N A is 24%, and single-copy 48%. W e have fixed
the reassociation rate constant o f the middle repetitive com-
p o n e n t to 1.93, and get a calculated complexity o f this com-
p o n e n t o f 1.25 x 10 s bp, a typical chloroplast figure for a
m e m b e r o f the m u s t a r d family (Lebacq a n d Vedel 1981).
The reassociation rate constant o f the single-copy c o m p o -
nent now becomes 0.0097, with a corresponding complexity
o f 5 x 107 bp. The r o o t mean square error becomes 0.015.
The calculated nuclear genome size from this curve, assum-
ing the middle repetitive c o m p o n e n t to be entirely chloro-
plast D N A , is 6.5 x 107 bp. The n u m b e r o f chloroplast ge-
nomes per h a p l o i d nuclear genome becomes 200. In other
words, letting the middle repetitive reassociation rate con-
stant float free or fixing it to a value that gives the middle
repetitive c o m p o n e n t a complexity exactly equal to that
expected o f chloroplast D N A makes little difference in our
calculations o f genome size or fraction o f genome in single-
copy D N A . Thus, under any set o f assumptions, or with
any reasonable constraints on the parameters o f our reasso-
ciation curves, we arrive at a p p r o x i m a t e l y the same estimate
o f the h a p l o i d genome size.
The measurements show that a p p r o x i m a t e l y 10% o f to-
tal Arabidopsis D N A is foldback, or highly repetitive. The
experiments reported shed no light on the nature o f this
D N A ; experiments addressing this question are in progress.
O u r calculations show, then, that the Arabidopsis nucle-
ar genome is c o m p a r a b l e in size to that o f the tiny n e m a t o d e
Caenorhabditis elegans (Sulston and Brenner 1974), less
than 20 times the size o f the Escherichia coli c h r o m o s o m e
(Cairns 1963), a n d only five times the size of the minute
Saccharomyces cerevisiae genome (Lauer et al. 1977). Fla-
vell (1980) has estimated that, if an average p l a n t messenger
is a b o u t 1,200 bases, and if there are a r o u n d 15,000 genes
per h a p l o i d genome in plants, the m i n i m u m a m o u n t o f
D N A r e q u i r e d for specifying the m R N A sequences o f a
typical p l a n t should be 1.8 x 107 NTP. Our kinetic measure-
m e n t o f the h a p l o i d Arabidopsis genome is a b o u t four times
this m i n i m u m figure. I f there are 15,000 genes in Arabidop-
sis, each occupies an average o f less than 5,000 base pairs
of chromosomal DNA.
The other characteristics o f the Arabidopsis genetic ma-
terial determined in this study, G + C content and 5-methyl-
cytosine levels, are not strikingly unusual, though the meth-
ylcytosine content is quite low for a higher plant. F u r t h e r
use o f r e c o m b i n a n t D N A techniques will soon show in
m o r e detail the nature o f the repetitive a n d methylated se-
quences. The future will also show if Arabidopsis thaliana,
with its unique c o m b i n a t i o n o f advantages for studies in
molecular genetics, will become an organism o f choice for
such studies.
Acknowledgements. We are very grateful to Drs. E. Diala and R.
Hoffman for their advice, aid, and the use of their equipment
in the 5-methylcytosine measurements, to Dr. R. Britten for the
use of his temperature-controlled spectrophotometer and for his
advice, to Drs. S. Johnson and J. Roberts for their advice, and
to Dr. E. Davidson for the loan of equipment and for helpful
suggestions. We thank Dr. A. Kleinhofs for Arabidopsisseeds. This
work was supported by grant number 82-CRCR-I-1063 from the
Science and Education Administration of the U.S. Department
of Agriculture to E.M.M.L.S.L. was supported by NIH Fellowship
1F32 GM 07725, and by the Gosney Fund.
References
Aerts M, Jacobs M, Hernalsteens J-P, van Montagu M, Schell
J (1979) Induction and in vitro culture of Arabidopsis thaliana
crown gall tumours. Plant Sci Lett 17:43-50
Bedbrook JR, Kolodner R (1979) The structure of chloroplast
DNA. Annu Rev Plant Physiol 30 : 593-620
Bennett MD, Smith JB (1976) Nuclear DNA amounts in angiosp-
erms. Proc R Soc Lond B 274:227-274
Bolivar E (1978) Construction and characterization of new cloning
vehicles. III. Derivatives of plasmid pBR322 carrying unique
EcoRI sites for selection of EcoRI generated recombinant DNA
molecules. Gene 4:121-136
Britten RJ, Graham DE, Neufeld BR (1974) Analysis of repeating
DNA sequences by reassociation. In: Grossman EL, Moldave
K (eds) Methods in enzymology, vol 29E. Academic Press, New
York, pp 363-406
Cairns J (1963) The chromosome of E. coll. Cold Spring Harbor
Symp Quant Biol 28 : 43-46
Crain WR, Eden FC, Pearson WR, Davidson EH, Britten RJ
(1976) Absence of short period interspersion of repetitive and
non-repetitive sequences in the DNA of Drosophila melano-
gaster. Chromosoma (Berl) 56:309-326
Diala E, Plant M, Coalson D, Hoffman R (1981) DNA methyla-
tion in normal and SV40-transformed human fibroblasts. Bio-
chem Biophys Res Commun 102:1379-1384
Diala E, Hoffman R (1982) Hypomethylation of HeLa cell DNA
and the absence of 5-methylcytosine in SV40 and adenovirus
(type 2) DNA: Analysis by HPLC. Biochem Biophys Res Com-
mun 107:19-26
Felsenfeld G (1971) Analysis of temperature-dependent absorption
spectra of nucleic acids. In: Cantoni GL, Davies DR (eds) Pro-
cedures in nucleic acid research, vol 2. Harper and Row Pub-
lishers, New York, pp 233-244
Flavell R (1980) The molecular characterization and organization
of plant chromosomal DNA sequences. Annu Rev Plant Phys-
iol 31 : 56%596
Gruenbaum Y, Navah-Many T, Cedar H, Razin A (1981) Sequence
specificity of methylation in higher plant DNA. Nature
292: 860-862
Kolodner R, Tewari KK (1975) The molecular size and conforma-
tion of the chloroplast DNA from higher plants. Biochim Bio-
phys Acta 402: 372-390
Koorneef M, van Eden J, Hanhart CJ, Stam P, Braaksma F J,
Feenstra WJ (1983) Linkage map of Arabidopsis thaliana. J
Hered 74:265-272

Kung S-D (1977) Expression of chloroplast genomes in higher
plants. Annu Rev Plant Physiol 28:401-437
Laird CD (1971) Chromatid structure: Relationship between DNA
content and nucleotide sequence diversity. Chromosoma (Berl)
32:378-406
Laird CD, McCarthy BJ (1969) Molecular characterization of the
Drosophila genome. Genetics 63 : 865-882
Lauer GD, Roberts TM, Klotz LC (1977) Determination of the
nuclear DNA content of Saecharomyces eerevisiae and implica-
tions for the organization of DNA in yeast chromosomes. J
Mol Biol 114:507-526
Lebacq P, Vedel F (1981) Sall restriction enzyme analysis of chlo-
roplast and mitochondrial DNAs in the genus Brassica. Plant
Sci Lett 23 : 1-9
Link G, Chambers SE, Thompson JA, Falk H (1981) Size and
physical organization of chloroplast DNA from mustard (Sina-
pis alba L.). Mol Gen Genet 181:454-457
Manning JE, Schmid CW, Davidson N (1975) Interspersion of
repetitive and nonrepetitive DNA sequences in the Drosophila
melanogaster genome. Cell 4:141-155
Pearson WR, Davidson EH, Britten RJ (1977) A program for
least squares analysis of reassociation and hybridization data.
Nucl Acids Res 4:1727-1737
R6dei GP (1970) Arabidopsis thaliana (L.) Heynh: A review of
the genetics and biology. Biblio Genetica 20:1-127
R6dei GP (1973) Extra chromosomal mutability determined by
a nuclear gene locus in Arabidopsis. Mutat Res 18:149-162
Rigby PWJ, Dieckmann M, Rhodes C, Berg P (1977) Labeling
deoxyribonucleic acid to high specific activity in vitro by nick
translation with DNA polymerase I. J Mol Biol 113:237-251
23
Schachat FH, Hogness DS (1973) Repetitive sequences in isolated
Thomas circles from Drosophila melanogaster. Cold Spring
Harbor Symp Quant Biol 38:371-381
Shapiro HS (1976) Distribution of purines and pyrimidines in de-
oxyribonucleic acids. CRC Handbook Biochem Molec Biol
Nucl Acids 2:241-281
Southern EM (1975) Detection of specific sequences among DNA
fragments separated by gel electrophoresis. J Mol Biol
98 : 503-517
Sulston JE, Brenner S (1974) The DNA of Caenorhabditis elegans.
Genetics 77:95 104
Thomas AJ, Sherrat HSA (1956) The isolation of nucleic acid frac-
tions from plant leaves and their purine and pyrimidine compo-
sition. Biochem J 62:1-4
Urieli-Shoval S, Gruenbaum Y, Sedat J, Razin A (1982) The ab-
sence of detectable methylated bases in Drosophila melanogaster
DNA. FEBS Lett 146:148-152
Vanyushin BF, Belozerskii AN (1959) Nucleotide composition of
the desoxyribonucleotides of higher plants. Dokl Akad Nauk
USSR 129:944-946
Vedel F, Mathieu C (1982) Isolation of purified mitochondrial
DNA from Brassicae. Anal Biochem 127:1-8
Communicated by R.B. Goldberg
Received October 16, 1983