| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 278, Issue 33, 30450-30457, August 15, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the
Departments of Biochemistry and Genetics,
Pediatrics, and
||Functional Genomics, Chiba University Graduate
School of Medicine, Chiba 260-8670, Japan, the
¶Department of Pharmacology and Brain Science,
School of Human Science, Waseda University, Tokorozawa 359-1192, Japan, and
the **Department of Pharmacology, Graduate School of
Pharmaceutical Sciences, Tohoku University, Sendai 980-8578, Japan
Received for publication, May 7, 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
-Hsd and Gabarapl1
genes exhibiting a peak in the light phase and Spot14, Hspa8, Hspa5,
and Hsp84-1 genes showing a peak in the dark phase. On the other
hand, mRNA levels for all of the three fibrinogen subunits, A
, B
and 
, exhibited two peaks each in the light and dark phases in a
synchronized manner. This two-peaked rhythmic pattern of fibrinogen genes as
well as the single peaktrough pattern of other genes was diminished or almost
completely lost in the liver of Clock mutant mice, suggesting that
the two-peaked expression is also under the control of oscillation-generating
genes. In constant darkness, the first peak of the expression rhythm of
fibrinogen genes was almost intact, but the second peak disappeared.
Therefore, although the first peak in the subjective day is a component of the
innate circadian rhythm, the second peak seems to require light stimuli.
Fasting in constant darkness caused shifts of time phases of the circadian
rhythms. Protein levels of the fibrinogen subunits in whole blood also
exhibited circadian rhythms. In the mouse and human loci of the fibrinogen
gene cluster, a number of sequence elements resembling circadian transcription
factor-binding sites were found. The fibrinogen gene locus provides a unique
system for the study of two-peaked day/night rhythms of gene expression in a
synchronized form. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
The time phase of the SCN oscillator is adjusted by light stimuli everyday in the light-dark condition (1, 2). The SCN then controls or affects the oscillators of peripheral cells in both relatively direct and indirect manners. The liver provides a well characterized example of these regulations. Although oscillation generators of liver cells are under the circadian control of glucocorticoids (9, 10), presumably through the hypothalamus-hypophysis-adrenal axis, time phases of the liver oscillators are more profoundly affected by feeding time (11–13). Artificial diurnal feeding of nocturnal rodents can completely uncouple the oscillation phase of the liver from that of the SCN. Seemingly, the SCN regulates the liver oscillators indirectly by controlling the sleep-awakeness phase that in turn determines the feeding time under the natural condition.
Recently, hundreds of genes exhibiting circadian oscillation in their mRNA
levels have been identified in the rodent liver
(14–19)
by comprehensive gene expression analysis including the microarray analysis,
showing that a number of liver functions are under the circadian control at
the gene expression level. These genes seem to show mainly circadian rhythms
with a single peak-trough a day, because most of the genes were identified by
the algorithms examining the fitness to cosine wave patterns with 
24 h of
cycles in constant darkness. Here, in the course of cDNA microarray analysis
and following RNA blot analysis, we found by chance that mRNAs for all of the
three fibrinogen subunits, A, B, and , exhibit peak levels
twice a day under the light-dark condition in the mouse liver. Fibrinogen, a
major component of blood coagulation, in the blood plasma has the subunit
composition of (A)2(B)22
from which N-terminal peptides A and B are proteolytically removed by
thrombin, yielding fibrin with the
222 composition
(20). Apparently,
corresponding with this stoichiometry of the subunit composition, the
two-peaked rhythms of fibrinogen genes were synchronized. In constant
darkness, the second peak of daily rhythms of fibrinogen genes disappeared,
whereas the first peak was almost intact. In human
(21) and rat
(22), the three fibrinogen
genes were shown to be clustered on chromosome 4 (4q31) and chromosome 2
(2q31–34), respectively, which is also the case for mouse chromosome 3
(3 E3-F1) as confirmed by the genome sequence
(23). Candidate sequence
elements for the rhythmic regulation were mapped in the mouse and human loci
of the fibrinogen gene cluster.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
cDNA Microarray Analysis—A microarray was prepared essentially as described previously (24) with 2,304 mouse liver cDNA clones provided by K. Hashimoto (National Institute of Infectious Diseases, Tokyo, Japan) and S. Sugano (Institute of Medical Sciences, University of Tokyo, Tokyo, Japan). Total liver RNA was prepared from ddY mice at ZT6 and 18 by the acid-guanidine thiocyanate-phenol/chloroform extraction procedure (25) and subjected to poly(A)+ RNA isolation using oligo(dT)-paramagnetic beads Dynabeads Oligo(dT)25 (Dynal, Oslo, Norway). 2 µg of poly(A)+ RNA and 4.5 µg of oligo(dT) in 15 µl of solution were heat-denatured at 70 °C for 10 min and immediately chilled on ice. The mixture was made up to 30 µl of the solution containing 50 mM Tris-HCl (pH 8.3), 75 mM KCl, 3 mM MgCl2, 10 mM DTT, 0.5 mM each of dATP, dCTP, and dGTP, 0.2 mM dTTP, 0.1 mM Cy3- or Cy5-dUTP, and 400 units of reverse transcriptase SuperScript II (Invitrogen). The reaction was allowed to proceed at 42 °C for 1 h. After alkaline degradation of template RNA, the Cy3- or Cy5-labeled cDNA was purified with Centricon-30 microconcentrators (Millipore, Bedford, MA). Hybridization was carried out in a solution containing 2 µg/µl yeast RNA, 2 µg/µl poly(A), 3.4x SSC (1x SSC consists of 0.15 M NaCl and 15 mM sodium citrate), and 0.3% SDS at 65 °C overnight under humidified condition. Washing was done twice for 5 min with 2x SSC, 0.1% SDS at room temperature and twice for 5 min with 0.2x SSC, 0.1% SDS at 40 °C followed by rinse with 0.2x SSC. The fluorescent images were scanned with a laser-scanning device (Scan-Array4000, GSI Lumonics, Bedford, MA).
Northern Blot Analysis—Total liver RNA and sense strand cRNA amplified from poly(A)+ RNA were subjected to Northern analysis. The detailed procedure for cRNA amplification will be described elsewhere. Double-stranded cDNA synthesized from poly(A)+ RNA absorbed on the oligo(dT) beads was ligated with the T7 promoter adaptor and heat-denatured, liberating the sense strand cDNA, which was again converted to double-stranded form priming from the oligo(dT) linker, amplified with PCR using as primers known sequences of both ends, and then subjected to synthesis of sense strand cRNA with T7 RNA polymerase. RNAs were electrophoresed in denaturing formaldehyde-agarose (1%) gels, visualized by ethidium bromide staining to check integrity and equal loading, and then blotted onto nylon membranes. Digoxigenin-labeled RNA probes were synthesized using a transcription kit (Roche Diagnostics) from cDNAs subcloned in the plasmid pGEM-3Zf(+). Hybridization, washing, and chemiluminescent detection on x-ray films were done as recommended by Roche Diagnostics. Densitometric quantification was performed by using Personal Scanning Image (PDSI, Molecular Dynamics, Sunnyvale, CA).
Western Blot Analysis—Whole blood was collected from the inferior vena cava of mice with heparinized syringes and tubes, and subjected to immunodetection of fibrinogen subunits essentially as described previously (26). The blood was serially diluted with 19 and 5.25 volumes of 70 mM Tris-HCl buffer (pH 6.8) containing 3% SDS, 5% 2-mercaptoethanol, 11.2% glycerol, and 0.02% bromphenol blue. The diluted samples containing 0.08 µl of whole blood in 20 µl of solution were subjected to SDS-PAGE with 10% acrylamide gel, and proteins were electrotransferred to nitrocellulose membranes. Immunodetection was performed using a goat IgG against mouse fibrinogen (1:1,000 dilution, Nordic Immunological Laboratories, Tilburg, The Netherlands) as a primary antibody, a peroxidase-conjugated rabbit IgG F(ab')2 against the goat IgG (1:5,000 dilution, Wako Pure Chemical Industries, Osaka, Japan) as a secondary antibody and ImmunoStar Reagents (Wako, Japan) for chemiluminescence detection. Signals detected on x-ray films were densitometrically quantified.
Data Base Search—cDNA sequences deposited in IMAGE were accessed through Search GenBankTM-Today Data Base using DBGET (www.genome.ad.jp/dbget-bin/www_bfind?genbank-today), queried in NCBI BLAST (www.ncbi.nlm.nih.gov:80/BLAST/), and assigned to specific genes with NCBI UniGene (www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=unigene) and NCBI LocusLink (www.ncbi.nlm.nih.gov/LocusLink/). Mouse and human genome sequences were from Mouse Genome Server (www.ensembl.org/Mus_musculus/) and Human Genome Server (www.ensembl.org/Homo_sapiens/), respectively, in Ensembl (www.ensembl.org/).
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
2,300
mouse liver cDNA clones. Poly(A)+ RNAs derived from the liver at
ZT6 and ZT18 in duplicate were subjected to Cy3 and Cy5 labeling,
respectively, coupled with cDNA synthesis. Both labeled cDNAs were mixed in an
equal amount and hybridized with a microarray. Labeling vice versa with Cy3
and Cy5 was also done to correct a possible difference in incorporation
efficiency of these dyes. Tables
I and
II represent lists for genes
that showed the higher mRNA levels at ZT6 and ZT18, respectively, at least
1.5-fold in all four experiments of the microarray analysis. Remarkably, genes
for all of the three fibrinogen subunits, A, B and ,
showed the higher expression at ZT6 (Table
I). It is also noteworthy that a number of genes possibly involved
in gene expression ranging from transcription to protein transport showed the
higher expression at ZT18 (Table
I). Spot14 is a putative transcriptional regulator
(27,
28), and leucine zipper
protein 1 is a basic protein localized in the nucleus
(29). Heat shock 70-kDa
protein 8 (Hspa8), heat shock 70-kDa protein 5 (Hspa5) (78-kDa
glucose-regulated protein), and heat shock protein 84-kDa 1 (Hsp84-1) are
molecular chaperones (30), and
peroxisome biogenesis factor 7 is involved in posttranslational protein
transport into the organelle
(31,
32). Seemingly, these genes
are involved in synthesis and delivery of proteins in a dark phase-specific
manner.
|
|
To confirm the validity of the microarray analysis, we performed Northern
analysis for six genes, i.e. two and four genes that showed the
higher expression at ZT6 and ZT18, respectively, more than 2-fold in all four
experiments. Total liver RNAs at ZT6 and ZT18 in duplicate were subjected to
the blot analysis (Fig.
1A). Concordant with the results of the microarray
analysis, mRNA levels for
3-hydroxy-
5-C27-steroid oxidoreductase
(3-HSD) (33) and GABA(A)
receptor-associated protein-like 1 (Gabarapl1)
(34) were higher at ZT6 than
ZT18 and those for Spot14, Hspa8, Hspa5, and Hsp84-1 were higher at ZT18 than
ZT6. We also performed Northern analysis by using poly(A)+
RNA-derived sense strand cRNA (Fig.
1B) to exclude the possibility that the results obtained
with total RNA may reflect daily fluctuation of poly(A)+ RNA
contents in total RNA. Again, changes in the gene expression levels detected
with the cRNA mixtures were concordant with the results of the microarray
analysis for both gene groups each exhibiting the higher expression at ZT6 or
ZT18. We concluded that our microarray system successfully revealed genes
differentially expressed between light and dark phases.
|
Genes Exhibiting Rhythmic Expression with a Single
Peak-Trough—Rhythmic patterns of expression of the six genes were
examined by Northern analysis with total RNA extracted from the liver of
wild-type mice at 4-h intervals in a light-dark cycle. In parallel, we also
investigated the mice with Clock mutation that yields a splice
variant of a dominant-negative type
(35) to see whether the
rhythmic expression of the six genes are under the control of
oscillation-generating genes. As shown in
Fig. 2A, mRNA levels
for 3-HSD and Gabarapl1 in the wild-type mice exhibited day/night rhythm
with a peak in the light phase and a trough in the dark phase. In the
Clock mutant mice, the rhythmic oscillations of these genes were
apparently attenuated, indicating that CLOCK plays a role in the regulation of
3-Hsd and Gabarapl1 genes.
|
Fig. 2B shows the results of similar analysis on Spot14, Hspa8, Hspa5, and Hsp84-1 genes with the higher expression in the dark phase. In the wild-type mice, each mRNA level for these genes showed a peak in the dark phase and a trough in the light phase. On the other hand, the rhythmic oscillations of the four genes were attenuated or almost completely lost in Clock mutant mice. Taken together with the results of Fig. 2A, both gene groups, each displaying a peak expression in the light or dark phase, seem to be under the control of the Clock gene.
Two-peaked Day/Night Rhythms in Expression of Fibrinogen
Genes—We also examined daily changes in expression of genes for
fibrinogen subunits by Northern analysis
(Fig. 3). Interestingly, in
wild-type mice, mRNA levels for all three fibrinogen subunits, A,
B, and , peaked twice at ZT7 and 19 in a day. The time phase and
oscillation amplitude of the two-peaked rhythm were synchronized among the
three genes. Since fibrinogen protein in the blood plasma consists of three
pairs of each subunit leading to the composition
(A)2(B)22
(20),
synchronization in expression of the subunit genes is suited for maintenance
of this stoichiometry. In Clock mutant mice, the rhythmicity appeared
to be shifted to the weak single peak-trough pattern for the A gene or
almost completely lost for the B and genes. Therefore, the
two-peaked expression of fibrinogen genes is also likely to be under the
control of the Clock gene.
|
To examine effects of the light-dark cycle and food intake on the daily rhythms of expression of fibrinogen genes, mice were kept in constant darkness and half of them were fasted (Fig. 4). Mice transferred to constant darkness on day 1 were subjected to liver excision at 4-h intervals from circadian time (CT) 3 to CT23 on day 2 (Fig. 4A). In Northern analysis for the fibrinogen subunit mRNAs (Fig. 4B, left panels), the second peak observed in the nighttime under the light-dark condition was not detected in the subjective night under constant darkness. On the other hand, the first peak was almost intact with some prolongation of the peak phase. Therefore, while the first peak in the subjective day is caused mainly by the innate circadian mechanism, the second peak seems to be generated by light stimuli (i.e. lights-on, continuous irradiation, and/or lights-off) most probably by lights-off at least as a component that affects the time phase of the second peak.
|
Effects of food depletion were also examined, because it is well known that feeding time profoundly affects the time phases of the liver oscillators (11–13). Foods were depleted at CT11 on day 1 just before beginning of the feeding period of nocturnal mice in the subjective night (Fig. 4A). Fasting caused shifts of time phases in the expression rhythms of fibrinogen subunit genes (Fig. 4B, right panels). Directions of the shifts remain to be determined experimentally, but we assume that forward shifts took place because fasting brings about forward shifts of the rhythms of mPer1 and mPer2 mRNA levels in the liver.2 Synchronization of the time phase among the subunit mRNAs seemed to be loosened by the fasting, the reason being remains to be clarified.
Rhythms of protein levels for the fibrinogen subunits in constant darkness were examined by Western analysis with whole blood (Fig. 4C). When mice were fed ad libitum (left panels), the three subunits exhibited synchronized circadian rhythms with an apparent peak at CT7 and possibly a low peak at CT15. Compared with the mRNA rhythms (Fig. 4B, left panels), this is rather an unexpected result. Direct reflection of mRNA levels to protein levels would have resulted in mimicking mRNA patterns by protein patterns with the lag time of several hours. Seemingly, circadian rhythms of fibrinogen protein levels are controlled not only by mRNA levels but also by other mechanisms such as regulations in the steps of translation, secretion, and protein degradation. Again, fasting appeared to cause shifts of time phases of fibrinogen protein rhythms, although individual variations of protein levels were relatively large under the fasting.
In human, it was demonstrated that plasma fibrinogen concentrations exhibit a day/night rhythm (36–38). Whereas the rhythm has been approximated by the cosine curve with a single peak in the early morning (38), reconsideration of the original data of Kanabrocki et al. (38) suggests the presence of the second peak or a shoulder in the evening. Therefore, human fibrinogen seems to be set to increase rapidly in the morning and persist around the daytime, i.e. a physically active period in human. The peak in the early morning has been implicated in the incidence of arterial ischemic diseases such as myocardial infarction (38).
Putative Regulatory Elements for Rhythmic Expression of the Fibrinogen
Gene Cluster—The two-peaked rhythmicity suggests that at least two
or possibly more transcription factors are involved in daily regulation of the
fibrinogen subunit genes. On the other hand, synchronization in expression of
subunit genes implies that they share common regulatory mechanisms. We
searched for putative cis-regulatory elements interacting with
circadian transcription factors in the fibrinogen gene cluster
(Fig. 5). Recent determination
of the mouse genome sequence
(23) confirmed the presence of
the fibrinogen gene cluster, which was predictable from the cytogenetic map
(Table I) and from previous
reports on homologous clusters of human
(21) and rat
(22). The searched sequences
were as follows: the E-box element CACGTG that is recognized by CLOCK/BMAL1
(3,
39) and NPAS2/BMAL1
(40) and that is also the
target of repressors such as PER/CRY
(5,
8) and recently characterized
Dec1 and Dec2 (41); the
REV-ERB site WAWNTRGGTCA (W = A or T, R = A or G) that is recognized by
members of nuclear orphan receptor REV-ERB and retinoic acid-related orphan
receptor families and that is responsible for circadian regulation
(19) by REV-ERB as a
well characterized member
(42); the DBP site RTTAYGTAAY
(R = AorG, Y = C or T) that is the binding sequence of circadian transcription
factors DBP, thyrotroph embryonic factor, and hepatocyte leukemia factor of
the PAR-ZIP (proline- and acidic amino acid-rich region-leucine zipper) family
(43) and its variant E4BP4
(44); and the glucocorticoid
response element (GRE) GGTACANNNTGTTCT
(45) that is the recognition
sequence of the nuclear receptor bound and activated by glucocorticoids, a
major humoral regulator of the liver clock
(9,
10). Completely matched
sequences for the E-box and one mismatch-allowed sequence for the other sites
are marked on mouse (Fig.
5A) and human (Fig.
5B) fibrinogen gene cluster loci with 10-kb flanking
regions.
|
A number of candidate sequences for each site were found. Exceptionally,
only one GRE-like sequence was detected in both mouse and human, and its
location is relatively similar between the two species, i.e. around
the 3'-terminal region of the fibrinogen B gene. Therefore, this
putative GRE may be functional and conservatory. Previous studies have
proposed the presence of GREs for rat A and B genes but have not
identified them (46). In the
5'-flanking region of the gene, sequences each resembling the
E-box, REV-ERB site, or DBP site seem to be arranged similarly between
mouse and human and may play regulatory roles common to both species. On the
other hand, frequencies of the three elements in the whole locus are rather
different between the two species and each frequency in the mouse
versus human locus is as follows (the number within parentheses shows
the frequency of completely matched sequences): the E-box 9
(9) versus 3
(3); the REV-ERB site 25
(4) versus 41
(2); and the DBP site 16 (0)
versus 44 (2). It is
tempting to speculate that relatively high frequencies of the E-box in mouse
and the REV-ERB and DBP sites in human are implicated in
species-specific regulations in mouse and human adapted to nocturnal and
diurnal activity, respectively. It will be extremely interesting to
investigate the roles of these putative cis-elements in two-peaked
synchronization in the expression of the fibrinogen gene cluster that is under
the control of both innate circadian mechanism and seemingly light-responsive
mechanism.
|
FOOTNOTES |
|---|
To whom correspondence should be addressed: Dept. of Biochemistry and
Genetics, Chiba University Graduate School of Medicine, Inohana 1-8-1, Chiba
260-8670, Japan. Tel.: 81-43-226-2035; Fax: 81-43-226-2037; E-mail:
mtak{at}med.m.chiba-u.ac.jp.
1 The abbreviations used are: SCN, suprachiasmatic nucleus; ZT, Zeitgeber
time; Hspa, heat shock 70-kDa protein; 3-HSD,
3-hydroxy-5-C27-steroid oxidoreductase
(dehydrogenase); Gabarapl1, GABA(A) receptor-associated protein-like 1;
Hsp84-1, heat shock protein 84-kDa 1; Fg, fibrinogen; CT, circadian time; GRE,
glucocorticoid response element; REV/ERB, reverse strand gene of
erbA; DBP, albumin site D-binding protein.
2 K. Horikawa, Y. Minami, M. Iijima, R. Aida, M. Akiyama, and S. Shibata,
manuscript in preparation.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Ishihara, E. Matsumoto, K. Horikawa, T. Kudo, E. Sakao, A. Nemoto, K. Iwase, H. Sugiyama, Y. Tamura, S. Shibata, et al. Multifactorial Regulation of Daily Rhythms in Expression of the Metabolically Responsive Gene Spot14 in the Mouse Liver J Biol Rhythms, August 1, 2007; 22(4): 324 - 334. [Abstract] [PDF]
|
|
|
|
 |
|
 D. F. Reilly, E. J. Westgate, and G. A. FitzGerald Peripheral Circadian Clocks in the Vasculature Arterioscler. Thromb. Vasc. Biol., August 1, 2007; 27(8): 1694 - 1705. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Bertolucci, M. Pinotti, I. Colognesi, A. Foa, F. Bernardi, and F. Portaluppi Circadian Rhythms in Mouse Blood Coagulation J Biol Rhythms, June 1, 2005; 20(3): 219 - 224. [Abstract] [PDF]
|
|
|
|
|
|
M. Pinotti, C. Bertolucci, F. Portaluppi, I. Colognesi, E. Frigato, A. Foa, and F. Bernardi Daily and Circadian Rhythms of Tissue Factor Pathway Inhibitor and Factor VII Activity Arterioscler. Thromb. Vasc. Biol., March 1, 2005; 25(3): 646 - 649. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
