Global Analysis of Functional Units of Plant Chromosomes:

1
Global Analysis of Functional Units of Plant
Chromosomes DNA Replication, Domain Structure,
and Transcription
PI Bill Thompsona, Co-PIs George Allenb, Linda
Hanley-Bowdoinc, Doreen Maind, Rob Martienssene,
Bryon Sosinskif, Matthew Vaughne
a Plant Biology, Genetics, and Crop Science, NC
State University, Raleigh, NC 27695 b
Horticultural Science, NC State University c
Biochemistry and Genetics, NC State University d
Horticulture Landscape Architecture, Washington
State University e Cold Spring Harbor
Laboratory, New York f Horticultural Science, NC
State University
Analysis of DNA Replication timing on Arabidopsis
Chromosome 4
In this project we are identifying and
characterizing regions of the genome that are
replicated at different times during S phase. We
have developed a FACS-based procedure, combined
with BrdU pulse-labeling and immunoprecipitation,
to analyze replication timing of an asynchronous
cell population. This approach is intended to
define functional domains of chromatin by
determining their preferred time of replication.
Figure 1 shows a chromosome-wide view of
genomic features and DNA replication in the
early, mid and late S-phase for Arabidopsis Chr
4. Panel A shows gene coverage (orange line) and
TE coverage (purple line). Panel B shows GC
percentage (calculated in 1-kb non-overlapping
windows) is shown in panel B. Panel C shows a
schematic representation of chr 4 omitting the
telomeres and NOR (The gene-rich euchromatic
distal short and distal long arms are shaded
light gray while the heterochromatic knob and
pericentromere are shaded black. The proximal
portions of both the short and long arms have
intermediate characteristics and are shaded dark
gray). Panel D shows replication profiles ,
expressed as log2 ratio of BrdU-incorporation in
early (blue), mid (green) or late (red) S-phase
cells relative to total DNA from the same cells.
Gene and TE coverage, GC percentage, and
replication profiles are loess-smoothed using a
150-kb window. Early replication is most
prevalent in the distal long arm, a predominately
euchromatic region. Late replication
predominates in the heterochromatic knob and
pericentromere, whereas regions of late
replication are dispersed in other portions of
the chromosome, especially in the
centromere-proximal portions of the long and
short arms. A remarkable feature of these
data is that the early and late replication
profiles in panel D show remarkable
complementarity (R -0.83), while the profiles
for replication in early and mid S-phase cells
are very similar to each other (R 0.87). The
most evident difference between the early and mid
S profiles is a broadening and merging of
early-replicating regions in mid S. In other
words, the DNA replicating in mid S-phase
represents nearly the same population of
sequences as that replicating in early S-phase.
The mid S-phase profile is also distinct from the
late profile (R -0.85). These data indicate
that replication in Arabidopsis is basically a
two phase process. The similarity of the early
and mid S profiles is best explained by assuming
considerable heterogeneity in the order of
replication in different cells. Figure 2
displays a schematic representation of replicons,
replication timing and replication domains for
chr 4. In the top panel, each vertical bar
represents a replicon, with the width of the bar
proportional to the length of the replicon.
Subdivisions within the bar indicate the
percentage of probes within the replicon with a
given replication time. The middle panel
illustrates the clustering of replicons with
similar timing into replication domains. The
lower panel is a cartoon of the major regions of
the chromosome, as in panel C of Figure 1. The
complexity of replication timing within many
replicons likely reflects several factors,
including time and efficiency of origin firing,
the number of origins within initiation zones,
and the speed of elongation by DNA polymerase in
specific contexts. Many of replication domains
we found in Arabidopsis chr 4 are considerably
smaller than those observed in mammalian cells.
However, several larger replication domains do
occur, including a 4.5-Mb late replication domain
at coordinates 2.6 7.1 Mb and a 2.3-Mb
early/mid replication domain located at
coordinates 16.2 18.5 Mb.
Mapping Nuclear Matrix Attachment Regions (MARs)

Genomic DNA is packaged and organized within the
nucleus by histones. When the histones are
extracted, the DNA forms large loops (nuclear
haloes), which remains bound by Matrix Attachment
Regions, or MARs, to a substructure composed of
RNA and protein called the nuclear matrix. While
the biological significance of MARs remains
largely unknown, several studies have shown that
MARs may function as origins of DNA replication
in higher eukaryotes. We have used lithium
diiodosalisylic acid (LIS) to extract the
histones from A. thaliana nuclei to produce
nuclear haloes, which were then digested with Eco
RI and Hind III. Matrix associated DNA was
separated from unbound DNA by low speed
centrifugation. The MAR DNA was then amplified
and labeled for microarray analysis.
Preliminary Mapping Results We have carried out
four array hybridization experiments that
represent two biological replicates and two
technical replicates using our custom-designed
NimbleGen tiling array for chromosome IV. Our
first pass analysis used a combination of limma
and NimbleScan, to resolve 933 putative MARs at
an estimated FDR lt0.05. The median length of the
putative MAR regions from this analysis is 800
bp. Panel A shows that the median AT content of
the putative MARs (histogram) is 71, which
contrasts to the median AT content of 63 along
chromosome IV (red curve). These data are
consistent with earlier studies showing that MARs
are AT-rich. The relative distance between each
MAR can be used to estimate loop size. Panel B
depicts the uneven spacing of MARs, which
encompass a range of loop sizes across chromosome
IV. The frequency of MAR spacing shows an
unimodal distribution. The largest peak contained
MARs with spacing that ranged from 42 bp to 265
kb with an average loop size of 19 kb and a
median loop domain size of 10 kb.
Mapping of Short Nascent Strands
DNA replication is a strictly regulated
process that preserves the genetic information
necessary for future generations. Despite its
importance, very little is known of the
regulation of DNA replication in higher
eukaryotes. Our goal is to understand and
define where DNA replication originates in the
Arabidopsis thaliana genome. We are using the
newly synthesized leading strands (short nascent
strands, SNS) which are thought to be initiated
at the very origins. These SNS are being analyzed
using a NimbleGen custom designed tiling array
that covers the entire chromosome lV. We have
developed two techniques to achieve this. In the
first (A), we isolated SNS by size using an
alkaline sucrose gradient (fig. A). Collected DNA
between 1 and 3kb (including SNS) is then
amplified, labeled and hybridized to the array.
The second technique was developed to enrich
and purify SNS. During the synthesis of DNA, the
RNA primase adds an RNA primer so the DNA
polymerase can recognize it and start
synthesizing DNA. To isolate SNS, we used lambda
exonuclease (B) which is unable to digest ssDNA
primed with RNA primer at 5 while digesting
unprotected DNA. This allows us to recover newly
synthesized DNA that is close to the origin. The
recovered SNS are then amplified, labeled and
hybridized to the array.
DNA Replication Timing in Rice
We are optimizing the essential conditions for
analysis of DNA replication timing on rice chr 4L
and 10L using a rice cell culture (cultivar
Nipponbare). All the technologies we developed
for Arabidopsis replication timing will apply to
this rice work with the optimized conditions.
Panel A The highest BrdU incorporation was
observed twelve to sixteen hours after 7 day
cultures were supplemented with fresh medium (7
day split). Panel B Analytical FACS profile of
1-hr pulse-labeled rice nuclei isolated from
cells after a 1 hour pulse given 16 hours after
7-day split cultures. Gates are defined for cell
populations in G1, S and G2/M phases.
Profiling Histone Modifications
During the last year we have profiled histone
modifications, RNA Pol II occupancy and gene
expression patterns in cell suspension culture
(samples Cells4 and Cells7, taken at 96hrs and 16
hrs post culture split, respectively) and in the
young rosette leaf samples from the wild-type Col
plants using Illumina GA2 high throughput DNA
sequencers. We used antibodies against histone
H3K4me2, H3K4me3, H3K9me2, H3K27me2, H3K27me3,
H3K14ac and H3K56ac, as well as antibodies
against RNA Pol II to immuno-precipitate DNA
associated with these histone modifications or
with the initiating or elongating forms of RNA
Pol II. This generated, on average, over 1.5x
genome coverage per sequenced ChIP library. We
have also sequenced over 2M mRNA fragments (40-50
nt in length) in a strand-specific manner from
the same samples. In addition, we have sequenced
between 1.7M and 3M small RNA from leaf, Cells4
and Cells7 samples. We are currently focusing our
efforts on the analysis of ChIP-seq and RNA-seq
data in the context of DNA replication timing by
computational identification of genomic regions
significantly enriched by immunoprecipitation
over the control, input, DNA, in 100bp windows.
The windows showing enrichment will be displayed
as intensity maps in the Generic Genome Browser 2
environment.
(C) Comparison of the preliminary data for SNS
with two different procedures. Data shows that
62 of the peaks found in SNS Lambda exonuclease
are common with SNS by size. While this suggest
that we have putative SNS, more analysis is
needed to confirm our results
ORC2 binding sites in Arabidopsis Ch4 euchromatin
often map with early replication timing peaks
Validation of ORC2 and MCM5 ChIP
(A) Proteins from the final ChIP input fraction
(40 ug, lane 1), whole cell extract (40 ug WCE,
lane 2) and volume equivalents from the
non-chromatin-associated (S, lane 3) and
chromatin-bound (P, lane 4) fractions were
resolved by SDS-PAGE, and the blots were probed
with the indicated antibodies. Chromatin was
immunoprecipitated with the indicated quantity of
anti-ORC2 serum (B) or anti- MCM5 serum (C).
Proteins remaining in the supernatant (depletion)
and the immunoprecipitated (enrichment) fractions
were resolved by SDS-PAGE and the blots were
probed with anti-ORC2 serum (B) or anti-MCM5
serum (C). (D) Sheared chromatin. Control (lane
2) and sonicated (lane 3) DNA was resolved and
visualized with ethidium bromide.
The profile shows the Loess-smoothed early S
replication profiles. Vertical lines indicate the
termination zones for putative replicons.ORC2-bind
ing sites are marked by orange spheres. Vertical
positioning of ORC2 sites is from the replication
profile and does not indicate ORC2 enrichment.
Preliminary analysis of ORC2 binding sites The
experiment included 3 bioreps each with an
IP-technical rep for a total of 6 ChIP samples
corresponding to ORC2. The samples were
hybridized to Nimblegene Ch4 tiling arrays. The
raw data was Loess-normalized and scaled in
limma, and peak-finding was performed with
NimbleScan using an 800-bp window. This analysis
identified 563 putative ORC2-binding sites at an
FDR 0.05, of which 289 were in constitutive
heterochromatin. Euchromatic ORC2 binding sites
tend to be intergenic, e.g. distal long arm is
only 22 intergenic but 56 of ORC2 binding
sites in this region are intergenic. ORC2 binding
sites are also AT-rich (68 for ORC2 binding
compared to 64 for sites not binding ORC2).