Upload
hanhi
View
232
Download
0
Embed Size (px)
Citation preview
1
A novel allele of kaiA shortens circadian period and strengthens interaction of oscillator 1
components in the cyanobacterium Synechococcus elongatus PCC 7942 2
3
You Chen,1†
Yong-Ick Kim,2†
Shannon R. Mackey,1†
C. Kay Holtman,1 Andy LiWang,
2† and 4
Susan S. Golden1†*
5
1Department of Biology, Texas A&M University, College Station, TX 77843-3258, USA 6
2Department of Biochemistry & Biophysics, Texas A&M University, College Station, TX 7
77843-2128, USA 8
9
Short running title: A Novel Short-Period kaiA Allele10
† Current address. Y Chen, Synthetic Genomics Inc., 11149 North Torrey Pines Road, La Jolla,
CA 92037; Y-I Kim & A LiWang, School of Natural Sciences, University of California at
Merced, 4225 N. Hospital Road, Atwater, CA 95301; SR Mackey, Biology Department, St.
Ambrose University, 518 West Locust Street, Davenport, IA 52803; SS Golden, Division of
Biological Sciences, University of California at San Diego, 9500 Gilman Drive, La Jolla, CA
92093, USA.
* Corresponding author. Mailing address: Division of Biological Sciences, University of
California at San Diego, 9500 Gilman Drive, La Jolla, CA 92093, USA. Phone: 858-246-0658.
Fax: 858-534-7108. E-mail: [email protected].
Copyright © 2009, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved.J. Bacteriol. doi:10.1128/JB.00334-09 JB Accepts, published online ahead of print on 24 April 2009
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
2
ABSTRACT 11
The basic circadian oscillator of the unicellular fresh water cyanobacterium Synechococcus 12
elongatus PCC 7942, the model organism for cyanobacterial circadian clocks, consists of only 13
three protein components: KaiA, KaiB and KaiC. These proteins, all of which are 14
homomultimers, periodically interact to form large protein complexes with stoichiometries that 15
depend on the phosphorylation state of KaiC. KaiA stimulates KaiC autophosphorylation 16
through direct physical interactions. Screening a library of S. elongatus transposon mutants for 17
circadian clock phenotypes uncovered an atypical short-period mutant that carries a kaiA 18
insertion. Genetic and biochemical analyses showed that the short-period phenotype is caused by 19
the truncation of KaiA by three amino acid residues at its C-terminus. The disruption of a 20
negative element upstream of the kaiBC promoter was another consequence of the insertion of 21
the transposon; when not associated with a truncated kaiA allele, this mutation extended the 22
circadian period. The circadian rhythm of KaiC phosphorylation was conserved in these mutants, 23
but with some modifications in the rhythmic pattern of KaiC phosphorylation, such as the ratio 24
of phosphorylated to unphosphorylated KaiC and the relative phase of the circadian 25
phosphorylation peak. The results showed that there is no correlation between the phasing of the 26
KaiC phosphorylation pattern and the rhythm of gene expression, measured as bioluminescence 27
from luciferase reporter genes. The interaction between KaiC and the truncated KaiA was 28
stronger than normal, as shown by fluorescence anisotropy analysis. Our data suggest that the 29
KaiA-KaiC interaction and the circadian pattern of KaiC autophosphorylation are both important 30
for determining the period, but not the relative phasing, of circadian rhythms in S. elongatus. 31
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
3
3
INTRODUCTION 32
The sun rises at dawn and sets at dusk with a predictable daily pattern. Many living organisms 33
show corresponding internal rhythms in their biological activities, including behavior, 34
metabolism and gene expression. These rhythms are generated by endogenous clocks, which 35
anticipate the daily environmental changes to optimize the timing of physiological activities of 36
an organism. Under constant conditions, these intrinsic rhythms show a self-sustained period of 37
about (circa) a day (dies), and are thus called circadian rhythms (4). They can be entrained to the 38
environmental cycle, like a clock being set to local time, through reception of light or 39
temperature cues at appropriate environmental phases. For circadian activities, the free-running 40
period of the rhythm is temperature compensated, that is, nearly constant over a physiological 41
range of ambient temperatures (11, 12). Diverse eukaryotic organisms, such as fungi, insects, 42
plants, birds and mammals including humans, have evolved to use this type of endogenous 43
biological clock (2, 46). Cyanobacteria, a group of photosynthetic eubacteria, are the simplest 44
organisms and the only prokaryotes known to possess a circadian clock (26). The majority of the 45
studies of cyanobacterial circadian rhythms have been carried out on using Synechococcus 46
elongatus PCC 7942, a unicellular fresh water obligate photoautotroph (8-10). 47
48
The kaiABC locus encodes the circadian pacemaker of S. elongatus. Both the monocistronic kaiA 49
transcript and the dicistronic kaiBC transcript display circadian cycling (15). The KaiB and KaiC 50
protein levels are also robustly rhythmic, whereas KaiA accumulation varies only modestly, or 51
imperceptibly, during the cycle (17, 20, 43). Mutations in any one of the kai genes cause varied 52
circadian period, ranging from 14 to 60 h (compared to 24-25 h in the wild type), or complete 53
arrhythmicity (27). The period length of cyanobacterial circadian rhythms has been proposed to 54
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
4
4
be determined by phosphorylation status and degradation rate of KaiC (33, 44). KaiA stimulates 55
autophosphorylation of KaiC; this stimulation is attenuated by KaiB (19, 23, 42, 44). The KaiA 56
dimer, KaiB tetramer, and KaiC hexamer compose large protein complexes in a circadian pattern, 57
which is closely associated with the phosphorylation status of KaiC (13, 20, 21, 30, 31, 45). The 58
circadian oscillation of KaiC phosphorylation can be regenerated in test tube in vitro simply by 59
mixing purified Kai proteins and ATP in appropriate ratio (33), and the circadian period can be 60
manipulated by changes in the relative concentrations of the Kai proteins (21). Despite recent 61
progress in elucidating the sequential phosphorylation of KaiC on Thr/Ser residues (35, 38) and 62
monomer exchange among KaiC hexamers (16), as well as the importance of these events in 63
circadian rhythm generation and synchronization, it is not fully understood how interactions 64
among Kai proteins determine the phosphorylation status of KaiC and how the phosphorylation 65
state is related to circadian period. 66
67
KaiA contains a pseudo-receiver domain in its amino terminal half that is proposed to receive 68
environmental cues, probably transmitted along circadian input pathways, and a carboxyl 69
terminal domain responsible for directly enhancing the autokinase activity of KaiC (19, 22, 23, 70
42, 44). A previous study showed that almost all kaiA point mutants displayed either arrhythmia 71
or a long-period phenotype (34). Here, we report a novel mutation in kaiA that causes an atypical 72
short-period phenotype. This insertional mutant was identified in a screen of an S. elongatus 73
transposon-mutant library that represents an insertion in essentially every gene and carries a 74
luciferase reporter gene to track circadian rhythms of bioluminescence (14). One mutant, which 75
has a shortened free-running period and elevated KaiB and KaiC protein levels, carries a 76
transposon insertion near the C-terminal end of the kaiA coding region. Further genetic and 77
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
5
5
biochemical analyses confirmed that the short period phenotype in this mutant was due to the 78
truncation of KaiA, and that the truncated KaiA variant interacts more strongly with the KaiC C-79
terminal peptide than does the wild-type KaiA protein. Elevation of KaiB and KaiC protein 80
accumulation resulted from disruption of a negative element upstream of the kaiBC promoter via 81
the insertion of the transposon. Changes in the circadian pattern of KaiC phosphorylation state 82
did not affect the phasing of bioluminescence rhythms. 83
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
6
6
MATERIALS AND METHODS 84
Cyanobacterial Strains, Media, and Culture Conditions. The cyanobacterial strains used in 85
this study are summarized in Table 1. All wild-type (WT) reporter and mutant strains were 86
created in Synechococcus elongatus PCC 7942. Cyanobacterial strains were grown in BG-11 87
medium (37) under continuous light (LL) conditions (~50-70 µE/m2s) at 30°C with appropriate 88
antibiotics. Two different versions of luciferase reporter genes were used: 1) firefly luciferase 89
(luc) in neutral site II (NS2); 2) Vibrio harveyi luciferase (luxAB) in the neutral site I (NS1) locus 90
and the aldehyde substrate synthesis genes luxCDE in NS2 (1). Neutral sites are segments of S. 91
elongatus genomic DNA, carried by cloning vectors, which mediate homologous recombination 92
with the S. elongatus chromosome and can be interrupted with no apparent phenotypes (1). 93
94
Plasmid Construction. Plasmids used in this report are described in Table 2. Unless otherwise 95
stated, plasmids were constructed in Escherichia coli strain DH10B (Invitrogen, Carlsbad, CA). 96
Plasmid pAM2246 (7) carries a copy of WT kaiA with its native promoter region in a NS1 vector. 97
The WT kaiA was then converted to kaiA281 using the QuickChange mutagenesis method (3) to 98
engineer a stop codon, which resulted in pAM3434. Plasmid pAM3582 was constructed by 99
inserting a PCR-amplified fragment, containing the original miniMuCm (GeneJumperTM
kit; 100
Invitrogen, Carlsbad, CA) transposon 7G3-GG5 (14) inserted at the 3’-end of the coding region 101
of kaiA along with its flanking sequences, into cloning vector pLitmus 29 (New England Biolabs, 102
Beverly, MA). The cat gene in the center of MuCm was then substituted by the gentamycin 103
resistance gene, aacA, from pAM3515 (pHP45ΩGm) to create pAM3613. The coding sequence 104
of kaiA was inserted into pET-32a(+) (Novagen, EMD, San Diego, CA) at HindIII/EcoRV sites 105
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
7
7
to build pAM3633. The last three amino acid residues of kaiA in pAM3633 were then deleted 106
using the QuickChange strategy to construct pAM3630. 107
108
Bioluminescence Assay and Data Analysis. The measurement of bioluminescence from the 109
reporter strains was performed as described previously (7, 14, 29). The acquired 110
bioluminescence data were processed and graphed by the Import and Analysis (IandA) Excel 111
macro set (S. A. Kay Laboratory, The Scripps Research Institute, La Jolla, CA). To calculate the 112
circadian periods of these strains, the data sets processed by IandA were then exported to 113
BRASS (Biological Rhythms Analysis Software System; A. J. Millar laboratory, University of 114
Edinburgh, Scotland, UK), which is an Excel interface for FFT-NLLS (a Fast Fourier Transform 115
statistical suite of programs; Dr. Martin Straume). The measurement of each strain was 116
conducted in at least three independent experiments with similar results. It was not technically 117
possible to assay all strains simultaneously, so strains were grouped to ensure that the most 118
relevant genotypes were assayed together; AMC541 was included with every run, and all data 119
from this strain were pooled to ensure reproducibility among assays (Table 3). For each strain 120
sample, the data of at least four circadian cycles in LL were used to calculate the period. 121
Standard error of the mean (SEM) was calculated in Excel (Microsoft, Redmond, WA). Unpaired 122
t-test that incorporates mean, SEM and N was performed to compare mutant or complemented 123
strains with corresponding WT strains. 124
125
Immunoblot Analysis. Protein sample preparation and immunoblot analysis for single time 126
point experiments were performed as describe previously (7) with slight modifications. To 127
facilitate sample collecting in a 24-h period, aliquots of cultures (O.D. 0.6-1.0 at 750 nm) for 128
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
8
8
time course analysis were first entrained in opposite phases at the same time with at least two 129
cycles of 12-h/12-h of either light/dark (LD) or dark/light (DL), and then released to LL 130
conditions (~50 µE/m2s). Samples were taken either every 4 h for 12 to 24 h or every 1 h for 4 h, 131
depending on the experimental design. 132
133
Peptide Purification and Fluorescein Labeling. Segments of the S. elongatus kaiC gene that 134
encode the desired peptide sequences were cloned in a pET-32a(+) vector (Novagen, EMD, San 135
Diego, CA), thereby creating thioredoxin–polyHis–peptide fusions (22, 36, 40). Escherichia coli 136
BL21(DE3) (Novagen, EMD, San Diego, CA) was transformed with the resulting plasmids and 137
grown in Luria broth. Expression of peptide fusion constructs was induced by adding isopropyl-138
β-D-thiogalactopyranoside (Calbiochem, EMD, San Diego, CA) to a final concentration of 1 139
mM. The cells were harvested after 4 h, resuspended in a buffer containing 50 mM NaCl and 20 140
mM Tris-HCl (pH 7.4), and lysed. Cell lysates were separated by centrifugation, and the fusion 141
peptide was purified from the supernatant fraction by metal affinity chromatography. The fusion 142
peptide was labeled with fluorescein at the N-terminus by the manufacturer’s protocol 143
(Molecular Probes, Eugene, OR). The sample was buffer exchanged to 50 mM NaCl, 20 mM 144
Tris-HCl (pH 7.4), and the peptide was cleaved from the affinity tag by using enterokinase 145
(Novagen, EMD, San Diego, CA). Cleavage by this enzyme results in the addition of three non-146
KaiC-derived residues (AMC) at the N-terminus of the peptide. Peptides were isolated by reverse 147
phase chromatography and lyophilized. Peptide identity and purity were confirmed by matrix-148
assisted laser desorption ionization time-of-flight mass spectroscopy, and quantification was 149
carried out by measuring the UV absorbance of fluorescein. 150
151
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
9
9
Fluorescence Anisotropy-Based Binding Experiments. Fluorescence anisotropy experiments 152
were carried out with an ISS PC1 photon counting spectrofluorometer (ISS, Champaign, IL) with 153
fluorescein-labeled KaiC peptide at 30ºC. The peptide concentration in all cases was fixed at 100 154
nM in 20 mM Tris-HCl, 150 mM NaCl, 0.5 mM EDTA (pH 8.0). An initial volume of 1.8 ml of 155
fluorescein-labeled KaiC peptide was used. Up to 2.2 ml of 0.3 mM KaiA-protein stocks were 156
added to the fluorescein-labeled KaiC peptide solution up to a total concentration of ~200 µM 157
protein. Fluorescence anisotropies were measured as a function of the concentrations of KaiA.158
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
10
10
RESULTS 159
Identification of an Atypical Short Period kaiA Mutant. Known kaiA mutants in S. elongatus 160
cause either arrhythmia or a long-period phenotype (34). When we screened mutants that carry 161
MuCm insertional mutations in a segment of the S. elongatus genome that is present on cosmid 162
7G3 for circadian clock phenotypes, we found an unusual short-period mutant allele of kaiA. The 163
insertion of MuCm (7G3-GG5) in kaiA shortened the free-running period length of the circadian 164
bioluminescence rhythm by about one h (AMC1483 & AMC1484, respectively, Fig. 1B & Table 165
3) in reporter strains that carry bacterial luciferase genes under the control of an S. elongatus 166
promoter: PpsbAI::luxAB in AMC1020 and PkaiBC::luxAB in AMC1300 (Table 1). Note that the 167
WT free-running period of the bioluminescence rhythm in AMC1300 is around 25 h, one h 168
longer than that of AMC1020 (~24 h) under the screening conditions used (see Discussion and 169
(5)). Regardless of the genetic background, the kaiA insertional mutant strains displayed the 170
same degree of period shortening and showed no discernible growth phenotype compared to their 171
corresponding WT strains. 172
173
Sequencing of the MuCm insertion site revealed its location at the C-terminal coding region of 174
kaiA. The insertion of MuCm separates the last 12 nucleotides, encoding the three terminal aa 175
residues RET plus the stop codon, from the rest of the coding region of kaiA (Fig. 1A). The end 176
sequence of the transposon forms a new stop codon for a truncated kaiA, which would encode a 177
protein with 281 aa instead of the 284 aa in WT KaiA. Immunoblot analysis of KaiA protein 178
levels confirmed the expression of a KaiA variant comparable to the WT KaiA protein in size 179
and amount, with slightly faster mobility on an SDS-PAGE gel (Fig. 1C). 180
181
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
11
11
There is a negative regulatory element that lies in the C-terminal-half of the kaiA coding region 182
and affects kaiBC expression (28) (Fig. 1A). The disruption of this element is known to 183
significantly increase the expression from the kaiBC promoter. Immunoblot analysis confirmed 184
elevated KaiB and KaiC abundance in the insertional mutant strains (Fig. 1C). A different kaiA 185
insertional knockout strain, AMC1161, was used as a negative control. This strain was generated 186
by inserting a kanamycin-resistant version of the Ω cassette into a BamHI site in the N-terminal 187
coding region of kaiA (7), such that the negative regulatory element of PkaiBC would not be 188
affected (Fig. 1A). As KaiA protein stimulates PkaiBC expression (15), disruption of kaiA in 189
AMC1161 caused a decrease in KaiB and KaiC protein levels such that only trace amount of 190
these two proteins could be detected (Fig. 1C). 191
192
In short, the 7G3-GG5 MuCm insertion not only truncated KaiA at its carboxyl terminus but also 193
disrupted a negative element and resulted in noticeably elevated levels of KaiB and KaiC protein 194
(Fig. 1C). Thus, the shortened period of the MuCm mutant could be due to either the truncation 195
of KaiA or the disruption of the negative element, or both. 196
197
The Short Period Phenotype Is Due to the Truncation of KaiA. To separate these two 198
possible causes for the clock phenotype, two different mutant strains were made, each mimicking 199
one of the two consequences of the MuCm insertion. Both are based on the AMC1161 200
background, in which the native kaiA locus carries a disrupting insertion in the N-terminal 201
coding region (kaiA null), and a copy of the firefly luciferase gene, under control of the kaiBC 202
promoter, serves as a reporter of gene expression. In one mutant, a truncated allele of kaiA 203
(kaiA281) was added ectopically and the negative element at the native kaiA locus remains intact. 204
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
12
12
In the other, the negative element was disrupted in the AMC1161 background at the same 205
position as the 7G3-GG5 transposon mutant, while a copy of WT kaiA was provided ectopically. 206
We predicted that the former would show the effect of truncated KaiA alone, whereas the latter 207
would recapitulate the effects of elevated KaiB and KaiC levels only. 208
209
As shown in Table 3 and Fig. 2A, the introduction of an ectopic copy of WT kaiA restores 210
circadian rhythmicity to the AMC1161 background (AMC1486) with a WT period of ~25 h. The 211
kaiA281 allele also restored rhythmicity to the kaiA null strain background (AMC1488), but with 212
a significantly shorter period (~24.5h). The protein levels of KaiA and KaiB in these two 213
complemented strains were not noticeably different from those of the WT strain (Fig. 2B). Thus, 214
the truncation of KaiA alone is not likely the cause of elevated KaiB and KaiC protein levels in 215
the original Mu mutant. The presence of two copies of WT kaiA shortened the period by almost 216
one h (AMC1485). The period was shortened even more when a copy of kaiA281 and a WT copy 217
of kaiA co-existed in the same strain (AMC1487, Table 3 & Fig. 2A). KaiA is known to 218
stimulate kaiBC expression (15). When two copies of kaiA/kaiA281 allele were present, we 219
indeed saw notably higher KaiB protein levels in these strains (Fig. 2B), which suggests that the 220
truncated KaiA281 protein retains the function to positively regulate kaiBC expression. In 221
conclusion, these results clearly show that the truncation of kaiA could cause a short period 222
phenotype, although it was not as marked as in the original transposon mutant strain. 223
224
Next we checked whether the disruption of the negative element also contributes to the shortened 225
period in the MuCm mutant strain. For compatibility of antibiotic resistance genes, the original 226
MuCm mutant was reconstructed with a gentamycin-resistance variant (MuGm). The insertion of 227
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
13
13
MuGm into kaiA at the same position as in the original MuCm strain shortened the period by 228
about 0.5 h in the PkaiBC::luc background (AMC1491). Cells remained arrhythmic after the 229
insertion of MuGm into kaiA in the kaiA insertional null background (AMC1492, Table 3 & Fig. 230
3A). The KaiB protein level was greatly elevated in the reconstructed MuGm mutant strain as 231
predicted for disruption of the negative element. In the absence of KaiA, only a trace amount of 232
KaiB was detected even when the negative element was also eliminated in the double-insertional 233
mutant strain (AMC1492, Fig. 3B). A copy of WT kaiA or kaiA281 was then introduced 234
ectopically into these two strains, respectively. In the double-insertional mutant strain 235
background, which has no KaiA and elevated KaiB and KaiC, the introduction of a copy of WT 236
kaiA restored the rhythmicity with a longer period, ~25.6 h (AMC1533, Fig. 3A & Table 3). 237
However, the introduction of a copy of kaiA281 in the same background, potentially mimicking 238
the original Mu insertional strains, could not restore a stable WT circadian rhythm in the cells 239
(AMC1534). The circadian traces of these cells showed reduced amplitude, advanced phase, and 240
fast damping phenotypes (Fig. 3A, the third pane). The period may also be shortened, but could 241
not be calculated accurately from the damped cycles. The addition of an ectopic copy of WT 242
kaiA did not significantly change the period in the AMC1491 background (kaiA281, disrupted 243
negative element), whereas an extra copy of kaiA281 in the same strain shortened the period 244
(AMC1531 and AMC1532, respectively, Fig. 3A & Table 3). Immunoblot data showed the 245
expected elevation in KaiB protein level when the negative element was disrupted in the 246
presence of an ectopic copy of KaiA or KaiA281 (Fig. 3B). Based on these results, we can 247
conclude that the disruption of the negative element along with normal expression level of KaiA 248
significantly elevates kaiBC expression, which results in an extended circadian period. 249
250
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
14
14
Together, these data indicate that the short period caused by the insertion of the Mu transposon in 251
kaiA is due to the truncation of KaiA. The kaiA281 allele, without disrupting basic KaiA function 252
as a central oscillator component, shortens the circadian period by ~0.5-1.0 h, depending on the 253
background of the reporter strain. On the contrary, disruption of the negative regulatory element 254
extends the period. Because the phosphorylation status of KaiC is thought to be the major factor 255
that determines the period length of cyanobacterial circadian clock (33, 44), we speculated that 256
the phosphorylation pattern of KaiC in the mutant strains would be different from that of WT 257
strains. 258
259
KaiC Autophosphorylation Pattern is Altered by KaiA281. A time course analysis was 260
performed to check the circadian pattern of KaiC phosphorylation in these strains. Synchronized 261
S. elongatus cultures were first sampled every 4 h under LL conditions (see Materials and 262
Methods). As shown in Fig. 4A, phosphorylated KaiC levels in the WT strain (AMC541) peaked 263
at around 8-12 h (LL8-12), while unphosphorylated KaiC peaked at LL0/24. Both kaiA- and 264
kaiA281-complemented strains (AMC1486 and AMC1488, respectively) displayed an increased 265
ratio of phosphorylated KaiC to unphosphorylated KaiC at LL0-4 and LL20-24. Samples taken 266
at 1-h intervals clearly showed that the level of phosphorylated KaiC in these strains was 267
elevated during these time points. In complemented strains a notable amount of phosphorylated 268
KaiC was seen even at LL0 and LL24, when mainly unphosphorylated KaiC was present in the 269
WT strain (Fig. 4B). However, the overall circadian pattern of unphosphorylated KaiC in either 270
complemented strains or the WT strain was very similar. There was also no difference in the 271
KaiC phosphorylation pattern between the kaiA281-complemented strain (AMC1488) and the 272
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
15
15
WT kaiA-complemented strain (AMC1486), even though they were slightly different in their 273
circadian period of bioluminescence. 274
275
We then checked the phosphorylation pattern of KaiC in strains in which the negative element 276
upstream of kaiBC is disrupted. As in the reconstructed MuGm mutant (AMC1491), the total 277
level of KaiC protein was elevated throughout the free-running circadian cycle in kaiA null 278
double-insertional mutant strains complemented with either kaiA (AMC1533) or kaiA281 279
(AMC1534) (Fig. 4C). AMC1491 and AMC1534 showed similar patterns, such that the elevated 280
KaiC was mainly in the unphosphorylated state. The circadian profile of accumulation of 281
phosphorylated KaiC was altered in these two strains compared to that of the WT strain: the rise 282
of phosphorylated KaiC was delayed at least 4 h. Although phosphorylated KaiC still peaked 283
around LL8-12 as in the WT strain, it then quickly diminished and was barely detectable at LL16, 284
when the WT strain still had abundant phosphorylated KaiC (Fig. 4C). The rise of 285
phosphorylated KaiC in AMC1533 was further delayed until LL12 and peaked at LL12-16 (Fig. 286
4C). It appeared that the KaiC phosphorylation cycle was shifted 4 h later in this strain compared 287
to the other strains. This observation was confirmed in samples taken in a continuous 24-h 288
sampling period (Fig. 4D). 289
290
From the results above, we can conclude that the truncation of KaiA does not significantly alter 291
its ability to stimulate autophosphorylation of KaiC, and that trans-complemented KaiA or 292
KaiA281 increases phosphorylated KaiC at times when it is low in WT strains (LL0-4 & LL20-293
24, Fig. 4A & B). On the other hand, the KaiC phosphorylation pattern in strains with elevated 294
KaiB and KaiC is different from the WT strain in some aspects, such as the phase of peak 295
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
16
16
phosphorylation. Nevertheless, the overall cycle of KaiC phosphorylation is retained. Thus, there 296
is no direct phase correlation between the phosphorylation pattern of KaiC and the 297
bioluminescence rhythms produced from the reporter genes. 298
299
The Interaction between KaiA and KaiC is Strengthened in KaiA281. The changes in the 300
phosphorylation pattern of KaiC and the period of bioluminescence cycles in mutant strains 301
suggest altered interactions between KaiA and KaiC. We performed fluorescence anisotropy-302
based binding assays to determine whether there is a change in the interaction between KaiA and 303
a KaiC-derived peptide, referred as CII ATP binding domain (CIIABD), in the KaiA281 variant. 304
CIIABD is the C-terminal peptide of KaiC and specifically binds KaiA (22, 36, 40). KaiA281 305
exhibited a tighter interaction with the KaiC peptide (Fig. 5) than did WT KaiA. The KaiA281 306
values were more similar to those of the C-terminal domain of KaiA, KaiA180C, which is 307
known to bind CIIABD more tightly and stimulate autokinase activity of KaiC more than does 308
full-length KaiA (22, 42). The results suggest that the short-period gene expression phenotype of 309
kaiA281 results from the strengthened KaiA-KaiC interaction.310
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
17
17
DISCUSSION 311
Among circadian clock period mutants, both short and long period alleles have been observed for 312
kaiC; however, only short period mutants have been identified in kaiB and usually long period 313
phenotypes result from mutation of kaiA (15, 34). This phenomenon is probably related to the 314
ability of the mutant proteins to affect the autokinase activity of KaiC, and the greater probability 315
of loss-of-function rather than gain-of-function mutants. A long period kaiA mutant, kaiA2 (also 316
called A30a), which carries a R249H missense point mutation, has a 30 h period (15). This 317
mutant has a reduced KaiC autokinase rate and accumulates unphosphorylated KaiC (19). Period 318
mutations in kaiA or kaiC have also been shown to alter the interaction between these two 319
proteins in yeast, even though the results were somewhat contradictory (39). 320
321
If there is a direct correlation between KaiC phosphorylation status and the circadian period, the 322
kaiA281 short period mutant might be expected to show more accumulated phosphorylated KaiC. 323
The time course analysis of KaiC protein samples, however, did not show notable differences in 324
accumulation of phosphorylated KaiC between a kaiA281-complemented kaiA insertional null 325
strain and its corresponding WT complementation strain (Fig. 4A & B). Because the period 326
difference between these two strains is less than 1 h, a change in KaiC phosphorylation status 327
might be present that is not revealed with the resolution of the time course. A KaiC autokinase 328
assay is a more quantitative way to check the difference between KaiA281 and WT KaiA in 329
stimulating KaiC autophosphorylation. Preliminary data showed that when KaiB was absent, the 330
sample containing purified KaiA281 did show slightly increased stimulation of KaiC 331
autophosphorylation activity compared with full-length KaiA (data not shown). These data are 332
consistent with the fact that long period KaiA mutants, including A30a/KaiA2, have a shorter 333
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
18
18
lifetime of phosphorylated KaiC, but do not differ much from WT strains in the turnover rate of 334
unphosphorylated KaiC (44). The prolonged KaiC phosphorylation state in the presence of 335
KaiA281 might also reflect differences in the degradation rate of KaiC in vivo. 336
337
The fluorescence anisotropy data indicate that the loss of the three KaiA C-terminal residues 338
alters the interaction between KaiA and KaiC (Fig. 5). KaiA binds specifically to the C-terminal 339
domain of KaiC possibly at two interfaces: CIIABD peptide and the ATP binding pocket (36, 40). 340
Both kaiA2 and kaiA281 mutants carry mutations that affect the C-terminal domain of KaiA, 341
which is responsible for interaction with KaiC and stimulation of KaiC autophosphorylation (41, 342
42). The R249H substitution encoded by kaiA2 is in the KaiA-KaiC interaction interface and 343
very close to the ATP binding pocket of KaiC; in contrast, the last three residues of KaiA map at 344
the edge of one of the dimer interfaces (45). Thus, deletion of these residues probably results in 345
minor conformational change of the KaiA dimer, which in turn strengthens the interaction 346
between KaiA and KaiC and enhances its ability to stimulate the autokinase activity of KaiC. 347
KaiA281 may not only interact more strongly with KaiC, but also affect the binding of KaiB to 348
KaiC and hence the dephosphorylation of KaiC. 349
350
The original MuCm kaiA insertion, 7G3-GG5, shortens the period in both PpsbAI::luxAB and 351
PkaiBC::luxAB reporter strains by almost 1 h. The period difference between the reconstructed 352
MuGm kaiA insertion mutant and its corresponding WT strain carrying a PkaiBC::luc reporter is 353
diminished to around 0.5 h. The two complemented kaiA strains, one with WT kaiA, and the 354
other with kaiA281, also show a period difference of ~0.5 h with each other (Table 3 & Fig. 6). 355
These differences may reflect the variations in the promoter activity and/or post-transcriptional 356
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
19
19
regulation of different reporter constructs. We have observed previously that the period length 357
differs among some WT reporter strains (5). 358
359
The short period kaiA281 allele does not lose its basic function as a central oscillator component. 360
Our results suggest that kaiA281 functions normally in enhancing kaiBC promoter activity and 361
sustaining the circadian oscillation (Table 3 and Fig. 2). Nevertheless, KaiA281 is different from 362
WT KaiA in some other aspects, as it could not completely complement AMC1492, in which 363
kaiA and the negative element are both disrupted by insertion, when provided ectopically (Fig. 364
3A). There might be a position effect of kaiA alleles (cis vs. trans); note that even the WT allele 365
did not complement with a normal period in this elevated KaiB/KaiC situation. Ectopic kaiA or 366
kaiA281 was introduced into the NS1 locus, which is located almost 180° from the native kai 367
locus on the circular chromosome. In all cyanobacteria species that carry kaiA, this gene is 368
always located immediate upstream of kaiBC. In addition, the negative element of PkaiBC is 369
located entirely within the coding region of kaiA in S. elongatus. This structural conservation of 370
the kai locus is likely to have significance in clock function. The ectopically expressed kaiA or 371
kaiA281 functions similarly to its cis counterpart in AMC1486, AMC1488, and AMC1533, in 372
which the oscillator is not otherwise compromised. It is known that a promoter with peak 373
expression 180° out of phase to that of PkaiBC, such as PpurF (7), or even the heterologous E. 374
coli promoter Ptrc expressed from a neutral site (32, 44), can successfully replace the original 375
kaiBC promoter to sustain a functional circadian clock in S. elongatus. Thus, there may be a 376
‘buffering mechanism’ in cyanobacterial clocks to tolerate certain internal alterations, which has 377
also been observed in a corresponding mammalian system (6). Circadian rhythms in 378
cyanobacteria can be sustained by either the phosphorylation oscillation or the transcriptional-379
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
20
20
translational feedback regulation of KaiC (24, 32). In the strain of AMC1534 that mimics the 380
original transposon insertional mutant, the combined effect of disrupting the negative element 381
and a truncated KaiA expressed in trans may simply compromise the timing mechanism beyond 382
its capacity for robustness. Although they show clear rhythmicity, complemented strains 383
AMC1486 and AMC1488 are different from the WT strain AMC541 in their circadian patterns 384
of KaiC phosphorylation. The increased amount of phosphorylated KaiC at the early and late 385
stages of a circadian cycle (Fig. 4A & B) may be another consequence of the ectopic expression 386
of kaiA or kaiA281. This pattern is not seen in AMC1533, probably because of the presence of 387
excess amount of unphosphorylated KaiC (Fig. 4C). 388
389
The ratio of KaiA to KaiC seems to be important to sustain a normal circadian period (21). Here 390
we show that when the KaiA to KaiC ratio is high due to an extra copy of kaiA, more KaiB (and 391
co-expressed KaiC that is not shown), accumulates and circadian period is shortened (as seen in 392
AMC1485 & AMC1487, Table 3 & Fig. 2B). Conversely, when the ratio is lowered due to the 393
disruption of the kaiBC upstream negative element, there is more KaiB (and KaiC) expressed 394
and a longer period (as seen in AMC1533, Table 3 & Fig. 3B). We propose that this correlation 395
is generally true in the cyanobacterial circadian clock. The kaiA281 allele also displays a 396
dominant effect: even with a WT kaiA allele present, a strain that encodes kaiA281 always has a 397
shorter circadian period than the WT strain on which it is based (Table 3 & Fig. 6). The 398
disruption of the negative element upstream of kaiBC promoter alone extends the period. In the 399
presence of a kaiA281 allele, however, the long period is masked by the short period phenotype, 400
as shown in the Mu insertional strains (Table 3). 401
402
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
21
21
No correlation between the peak time of accumulation of phosphorylated KaiC and rhythmic 403
bioluminescence traces can be deduced from our results. Strain AMC1534, mimicking the 404
original 7G3-GG5 MuCm insertional strain, is phase advanced in its bioluminescence rhythms, 405
but not in KaiC phosphorylation pattern. On the contrary, AMC1533 shows a normal phase in 406
bioluminescence traces, but is phase delayed (~4 h) in time course immunoblot analysis (Fig. 4C 407
& D). Previously, a point mutation of kaiC (C1265T) was shown to cause the loss of the 408
rhythmic KaiC accumulation and reduced circadian oscillation in phosphorylation. However, this 409
mutant, called pr1, still possesses robust bioluminescence rhythms with a WT period, albeit 3-4 410
h phase advanced (25). Thus, there must be other mechanisms for phase regulation, probably 411
through a circadian output pathway. Nonetheless, the overall circadian rhythms of the KaiC 412
phosphorylation cycles are conserved in these strains. Because the period difference among all 413
the strains in this study (except AMC1534) is less than 1 h, the KaiC phosphorylation and 414
dephosphorylation cycle recurs after approximately 24 h despite variations in phosphorylation 415
pattern during the cycle.416
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
22
22
ACKNOWLEDGMENTS 417
We thank Larry Harris Haller and the Gene Technologies Laboratory staff for technical advice 418
and support. 419
420
This work was supported by grants from the National Science Foundation (MCB-9818031) 421
(SSG), the National Institute of Neurological Disorders and Stroke (P01 NS39546) (SSG), and 422
the National Institutes of Health (GM064576) (AL).423
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
23
23
REFERENCES 424
1. Andersson, C. R., N. F. Tsinoremas, J. Shelton, N. V. Lebedeva, J. Yarrow, H. Min, 425
and S. S. Golden. 2000. Application of bioluminescence to the study of circadian 426
rhythms in cyanobacteria. Methods Enzymol. 305:527-42. 427
2. Bell-Pedersen, D., V. M. Cassone, D. J. Earnest, S. S. Golden, P. E. Hardin, T. L. 428
Thomas, and M. J. Zoran. 2005. Circadian rhythms from multiple oscillators: lessons 429
from diverse organisms. Nat. Rev. Genet. 6:544-56. 430
3. Braman, J., C. Papworth, and A. Greener. 1996. Site-directed mutagenesis using 431
double-stranded plasmid DNA templates. Methods Mol. Biol. 57:31-44. 432
4. Bünning, E. 1973. The Physiological Clock: Circadian Rhythms and Biological 433
Chronometry. Springer, New York. 434
5. Clerico, E. M., V. M. Cassone, and S. S. Golden. 2009. Stability and lability of 435
circadian period of gene expression in the cyanobacterium Synechococcus elongatus. 436
Microbiology 155:635-41. 437
6. Dibner, C., D. Sage, M. Unser, C. Bauer, T. d'Eysmond, F. Naef, and U. Schibler. 438
2009. Circadian gene expression is resilient to large fluctuations in overall transcription 439
rates. EMBO J 28:123-34. 440
7. Ditty, J. L., S. R. Canales, B. E. Anderson, S. B. Williams, and S. S. Golden. 2005. 441
Stability of the Synechococcus elongatus PCC 7942 circadian clock under directed anti-442
phase expression of the kai genes. Microbiology 151:2605-13. 443
8. Ditty, J. L., S. B. Williams, and S. S. Golden. 2003. A cyanobacterial circadian timing 444
mechanism. Annu. Rev. Genet. 37:513-43. 445
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
24
24
9. Golden, S. S. 2003. Timekeeping in bacteria: the cyanobacterial circadian clock. Curr. 446
Opin. Microbiol. 6:535-40. 447
10. Golden, S. S., and S. R. Canales. 2003. Cyanobacterial circadian clocks--timing is 448
everything. Nat. Rev. Microbiol. 1:191-9. 449
11. Golden, S. S., C. H. Johnson, and T. Kondo. 1998. The cyanobacterial circadian 450
system: a clock apart. Curr. Opin. Microbiol. 1:669-73. 451
12. Harmer, S. L., S. Panda, and S. A. Kay. 2001. Molecular bases of circadian rhythms. 452
Annu. Rev. Cell Dev. Biol. 17:215-53. 453
13. Hitomi, K., T. Oyama, S. Han, A. S. Arvai, and E. D. Getzoff. 2005. Tetrameric 454
architecture of the circadian clock protein KaiB. A novel interface for intermolecular 455
interactions and its impact on the circadian rhythm. J. Biol. Chem. 280:19127-35. 456
14. Holtman, C. K., Y. Chen, P. Sandoval, A. Gonzales, M. S. Nalty, T. L. Thomas, P. 457
Youderian, and S. S. Golden. 2005. High-Throughput Functional Analysis of the 458
Synechococcus elongatus PCC 7942 Genome. DNA Res. 12:103-15. 459
15. Ishiura, M., S. Kutsuna, S. Aoki, H. Iwasaki, C. R. Andersson, A. Tanabe, S. S. 460
Golden, C. H. Johnson, and T. Kondo. 1998. Expression of a gene cluster kaiABC as a 461
circadian feedback process in cyanobacteria. Science 281:1519-23. 462
16. Ito, H., H. Kageyama, M. Mutsuda, M. Nakajima, T. Oyama, and T. Kondo. 2007. 463
Autonomous synchronization of the circadian KaiC phosphorylation rhythm. Nat Struct 464
Mol Biol 14:1084-1088. 465
17. Ivleva, N. B., M. R. Bramlett, P. A. Lindahl, and S. S. Golden. 2005. LdpA: a 466
component of the circadian clock senses redox state of the cell. EMBO J. 24:1202-10. 467
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
25
25
18. Iwasaki, H., and T. Kondo. 2004. Circadian timing mechanism in the prokaryotic clock 468
system of cyanobacteria. J. Biol. Rhythms. 19:436-44. 469
19. Iwasaki, H., T. Nishiwaki, Y. Kitayama, M. Nakajima, and T. Kondo. 2002. KaiA-470
stimulated KaiC phosphorylation in circadian timing loops in cyanobacteria. Proc. Natl. 471
Acad. Sci. USA 99:15788-93. 472
20. Kageyama, H., T. Kondo, and H. Iwasaki. 2003. Circadian formation of clock protein 473
complexes by KaiA, KaiB, KaiC, and SasA in cyanobacteria. J. Biol. Chem. 278:2388-95. 474
21. Kageyama, H., T. Nishiwaki, M. Nakajima, H. Iwasaki, T. Oyama, and T. Kondo. 475
2006. Cyanobacterial circadian pacemaker: Kai protein complex dynamics in the KaiC 476
phosphorylation cycle in vitro. Mol. Cell. 23:161-71. 477
22. Kim, Y. I., G. Dong, C. W. Carruthers, Jr., S. S. Golden, and A. LiWang. 2008. The 478
day/night switch in KaiC, a central oscillator component of the circadian clock of 479
cyanobacteria. Proc Natl Acad Sci U S A 105:12825-30. 480
23. Kitayama, Y., H. Iwasaki, T. Nishiwaki, and T. Kondo. 2003. KaiB functions as an 481
attenuator of KaiC phosphorylation in the cyanobacterial circadian clock system. EMBO 482
J. 22:2127-34. 483
24. Kitayama, Y., T. Nishiwaki, K. Terauchi, and T. Kondo. 2008. Dual KaiC-based 484
oscillations constitute the circadian system of cyanobacteria. Genes Dev 22:1513-21. 485
25. Kiyohara, Y. B., M. Katayama, and T. Kondo. 2005. A novel mutation in kaiC affects 486
resetting of the cyanobacterial circadian clock. J. Bacteriol. 187:2559-64. 487
26. Kondo, T., C. A. Strayer, R. D. Kulkarni, W. Taylor, M. Ishiura, S. S. Golden, and 488
C. H. Johnson. 1993. Circadian rhythms in prokaryotes: luciferase as a reporter of 489
circadian gene expression in cyanobacteria. Proc. Natl. Acad. Sci. USA 90:5672-6. 490
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
26
26
27. Kondo, T., N. F. Tsinoremas, S. S. Golden, C. H. Johnson, S. Kutsuna, and M. 491
Ishiura. 1994. Circadian clock mutants of cyanobacteria. Science 266:1233-6. 492
28. Kutsuna, S., Y. Nakahira, M. Katayama, M. Ishiura, and T. Kondo. 2005. 493
Transcriptional regulation of the circadian clock operon kaiBC by upstream regions in 494
cyanobacteria. Mol. Microbiol. 57:1474-84. 495
29. Mackey, S. R., J. L. Ditty, E. M. Clerico, and S. S. Golden. 2007. Detection of 496
rhythmic bioluminescence from luciferase reporters in cyanobacteria. Methods Mol. Biol. 497
362:115-29. 498
30. Mori, T., S. V. Saveliev, Y. Xu, W. F. Stafford, M. M. Cox, R. B. Inman, and C. H. 499
Johnson. 2002. Circadian clock protein KaiC forms ATP-dependent hexameric rings and 500
binds DNA. Proc. Natl. Acad. Sci. USA 99:17203-8. 501
31. Mori, T., D. R. Williams, M. O. Byrne, X. Qin, M. Egli, H. S. McHaourab, P. L. 502
Stewart, and C. H. Johnson. 2007. Elucidating the Ticking of an In Vitro Circadian 503
Clockwork. PLoS Biol. 5:e93. 504
32. Murayama, Y., T. Oyama, and T. Kondo. 2008. Regulation of circadian clock gene 505
expression by phosphorylation states of KaiC in cyanobacteria. J Bacteriol 190:1691-8. 506
33. Nakajima, M., K. Imai, H. Ito, T. Nishiwaki, Y. Murayama, H. Iwasaki, T. Oyama, 507
and T. Kondo. 2005. Reconstitution of circadian oscillation of cyanobacterial KaiC 508
phosphorylation in vitro. Science 308:414-5. 509
34. Nishimura, H., Y. Nakahira, K. Imai, A. Tsuruhara, H. Kondo, H. Hayashi, M. 510
Hirai, H. Saito, and T. Kondo. 2002. Mutations in KaiA, a clock protein, extend the 511
period of circadian rhythm in the cyanobacterium Synechococcus elongatus PCC 7942. 512
Microbiology 148:2903-9. 513
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
27
27
35. Nishiwaki, T., Y. Satomi, Y. Kitayama, K. Terauchi, R. Kiyohara, T. Takao, and T. 514
Kondo. 2007. A sequential program of dual phosphorylation of KaiC as a basis for 515
circadian rhythm in cyanobacteria. Embo J 26:4029-37. 516
36. Pattanayek, R., D. R. Williams, S. Pattanayek, Y. Xu, T. Mori, C. H. Johnson, P. L. 517
Stewart, and M. Egli. 2006. Analysis of KaiA-KaiC protein interactions in the cyano-518
bacterial circadian clock using hybrid structural methods. EMBO J. 25:2017-28. 519
37. Rippka, R., J. Deruelles, J. Waterbury, M. Herdman, and R. Stanier. 1979. Generic 520
assignments, strain histories and properties of pure cultures of cyanobacteria. J. Gen. 521
Microbiol. 111:1-61. 522
38. Rust, M. J., J. S. Markson, W. S. Lane, D. S. Fisher, and E. K. O'Shea. 2007. 523
Ordered phosphorylation governs oscillation of a three-protein circadian clock. Science 524
318:809-12. 525
39. Taniguchi, Y., A. Yamaguchi, A. Hijikata, H. Iwasaki, K. Kamagata, M. Ishiura, M. 526
Go, and T. Kondo. 2001. Two KaiA-binding domains of cyanobacterial circadian clock 527
protein KaiC. FEBS Lett. 496:86-90. 528
40. Vakonakis, I., and A. C. LiWang. 2004. Structure of the C-terminal domain of the 529
clock protein KaiA in complex with a KaiC-derived peptide: implications for KaiC 530
regulation. Proc. Natl. Acad. Sci. USA 101:10925-30. 531
41. Vakonakis, I., J. Sun, T. Wu, A. Holzenburg, S. S. Golden, and A. C. LiWang. 2004. 532
NMR structure of the KaiC-interacting C-terminal domain of KaiA, a circadian clock 533
protein: implications for KaiA-KaiC interaction. Proc. Natl. Acad. Sci. USA 101:1479-84. 534
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
28
28
42. Williams, S. B., I. Vakonakis, S. S. Golden, and A. C. LiWang. 2002. Structure and 535
function from the circadian clock protein KaiA of Synechococcus elongatus: a potential 536
clock input mechanism. Proc. Natl. Acad. Sci. USA 99:15357-62. 537
43. Xu, Y., T. Mori, and C. H. Johnson. 2000. Circadian clock-protein expression in 538
cyanobacteria: rhythms and phase setting. EMBO J. 19:3349-57. 539
44. Xu, Y., T. Mori, and C. H. Johnson. 2003. Cyanobacterial circadian clockwork: roles of 540
KaiA, KaiB and the kaiBC promoter in regulating KaiC. EMBO J. 22:2117-26. 541
45. Ye, S., I. Vakonakis, T. R. Ioerger, A. C. LiWang, and J. C. Sacchettini. 2004. 542
Crystal structure of circadian clock protein KaiA from Synechococcus elongatus. J. Biol. 543
Chem. 279:20511-8. 544
46. Young, M. W., and S. A. Kay. 2001. Time zones: a comparative genetics of circadian 545
clocks. Nat. Rev. Genet. 2:702-15. 546
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
29
29
FIGURE LEGENDS 547
FIG. 1. An atypical short-period kaiA insertional mutant. A) Schematic representation of the kai 548
locus showing the relative positions of insertions: an Omega Cassette (ΩKm) and MuCm 549
transposon 7G3-GG5 (Mu). ΩKm was inserted in the N-terminal coding region of kaiA to create 550
AMC1161, a kaiA null. The Mu mutant encodes KaiA281, truncated by 3 residues, and disrupts a 551
negative control element upstream of the kaiBC promoter. A bent arrow shows the position of 552
the kaiBC promoter. The dashed line under the C-terminal coding region of kaiA indicates the 553
kaiBC upstream negative element. B) Representative bioluminescence traces from WT (closed 554
symbols) and the Mu mutant (open symbols) in two different reporter backgrounds: 555
PpsbAI::luxAB (AMC1020, squares) and PkaiBC::luxAB (AMC1300, triangles). Peaks of 556
circadian rhythms are indicated by vertical lines. X-axis, time in h: the blank bars represent light 557
conditions; the black bar represents 12-h darkness. Y-axis, counts per second, cps. C) 558
Immunoblot analysis of Kai proteins in WT and Mu-insertional mutant strains. Total soluble 559
protein (~20 µg) was loaded in each well. AMC702 and AMC705 carry in-frame deletions of 560
kaiA (∆A) or kaiBC (∆BC), respectively; AMC1161 carries an insertional null allele of kaiA 561
(∇A); WT, the wild-type strain; Mu stands for the 7G3-GG5 MuCm insertion strains (AMC1483 562
& AMC1484). ns, not shown. 563
564
FIG. 2. Truncation of KaiA by 3 residues at its carboxyl terminus causes a short period. A) 565
Representative bioluminescence traces from the WT strain (AMC541, closed circles), the 566
insertional kaiA null strain (AMC1161, dashed lines), and complemented strains (open symbols) 567
in the reporter background of PkaiBC::luc. The WT and null strains are shown only in the top 568
panel. Open circles, an ectopic copy of WT kaiA in AMC1161 background (AMC1486); open 569
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
30
30
triangles, an ectopic copy of truncated kaiA (kaiA281) in AMC1161 (AMC1488); open diamonds, 570
an ectopic copy of WT kaiA in AMC541 (AMC1485); open squares, an ectopic copy of kaiA281 571
in AMC541 (AMC1487). The last peak of WT circadian rhythms is indicated by a vertical dotted 572
line, while that of complemented strains is positioned with a vertical black line. Axes are as for 573
Fig. 1. B) Immunoblot analysis of KaiA and KaiB proteins performed as for Fig. 1. KaiC 574
consistently behaved like KaiB and is not shown here. AMC702, AMC705, and AMC1161 are 575
described in Fig. 1. WT, wild-type kaiA; WW, and 2W, an ectopic kaiA or kaiA281 allele, 576
respectively, in a WT background; W∇A and 2∇A indicate an ectopic wild-type kaiA or kaiA281 577
allele, respectively, in a kaiA insertional null background. ns, not shown. 578
579
FIG. 3. Disruption of the upstream negative element of the kaiBC promoter increases KaiB 580
protein level and slows the clock. A) Representative bioluminescence traces from the 581
PkaiBC::luc reporter background. The following two reference strains are shown only in the top 582
panel: closed circles, WT background (AMC541); short horizontal bars, kaiBC negative element 583
disrupted and insertional kaiA null (AMC1492). From the top pane: open circles, reconstructed 584
MuGm mutant (AMC1491); open diamonds (AMC1533) and long horizontal bars (AMC1534) 585
carry an ectopic WT and a kaiA281 allele of kaiA, respectively, in an AMC1492 background; 586
open squares (AMC1531) and open triangles (AMC1532) carry an ectopic WT and a kaiA281 587
allele of kaiA, respectively, in an AMC1491 background. The last peak of WT circadian rhythms 588
is indicated by a vertical dotted line, while that of complemented strains is positioned with a 589
vertical black line. There is an exception in the third pane, where circadian rhythms damped after 590
two cycles in LL in the complemented strain, AMC1534. Instead, two vertical lines were drawn 591
to indicate the second peak in LL for AMC541 and AMC1534. Axes are as described for Fig. 1. 592
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
31
31
B) Immunoblot analysis of KaiA and KaiB proteins performed as for Fig. 1. AMC541, AMC702, 593
AMC705, and AMC1161 strains are described in Fig. 1 & 2. WT, wild-type kaiA; MG, MuGm in 594
a WT background; WM and 2M indicate an ectopic kaiA or kaiA281 allele in a MuGm 595
background, respectively; M∇A, MuGm in the kaiA insertional null (AMC1161) background; 596
WM∇ and 2M∇, an ectopic WT or kaiA281 allele in the M∇A background, respectively. ns, not 597
shown. 598
599
FIG. 4. KaiC autophosphorylation pattern in cyanobacterial strains. A) Time course immunoblot 600
analysis of S. elongatus strains AMC541 (WT), AMC1486 (WT kaiA-complemented strain) and 601
AMC1488 (kaiA281-complemented strain). 4-h time point samples taken over 12 h. B) 1-h time 602
point samples taken over 4 h for strains in A). C) Time course immunoblot analysis of S. 603
elongatus strains AMC541, AMC1491 (MuGm), AMC1533 (WT kaiA-complemented and 604
negative element disrupted strain) and AMC1534 (kaiA281-complemented and negative element 605
disrupted strain). 4-h time point samples taken over 12 h. D) 4-h time point samples taken over 606
24 h for AMC1533. LD (light/dark) and DL (dark/light) indicate opposite entrainment conditions. 607
The hours released into LL for each time point is shown on top of each lane. P-kaiC-, 608
phosphorylated KaiC; KaiC-, unphosphorylated KaiC. Total soluble protein (~40 µg) was loaded 609
in each well. 610
611
FIG. 5. Fluorescence anisotropy data of 6-iodoacetamido-fluorescein-labeled KaiC peptides as a 612
function of KaiA180C (squares), WT KaiA (triangles), and KaiA281 (circles). The KaiC peptide 613
consists of the C-terminal 32 residues of S. elongatus KaiC. All fluorescence anisotropy values 614
were measured at 30ºC. 615
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
32
32
616
FIG. 6. Column chart of circadian period of cyanobacterial strains in Table 1. X-axis, 617
cyanobacterial strains; Y-axis, circadian period in h. *P < 0.0001, unpaired t-test compared with 618
corresponding WT strains. 619
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
Table 1. Cyanobacterial strains Strain # Genetic
background
Ectopic kaiA Reporter gene (s) * Source or
reference
AMC541 Wild-type - PkaiBC::luc
(pAM2105, NS2, CmR)
Ditty et al.
(2003)
AMC702 kaiA in-frame
deletion
- PkaiBC::luc
(pAM2105, NS2, CmR)
Ditty et al.
(2005)
AMC705 kaiBC in-frame
deletion
- PkaiBC::luc
(pAM2105, NS2, CmR)
Ditty et al.
(2005)
AMC1020 Wild-type - PpsbAI::luxAB
(pAM1501, NS1, SpR);
PpsbAI::luxCDE
(pAM1619, NS2, KmR)
Andersson et al.
(2000)
AMC1161 kaiA ΩKm
insertion
- PkaiBC::luc
(pAM2105, NS2, CmR)
Ditty et al.
(2005)
AMC1300 Wild-type - PkaiBC::luxAB
(pAM1887, NS1, SpR);
PpsbAI::luxCDE
(pAM1619, NS2, KmR)
Lab collection
AMC1483 kaiA MuCm
insertion
- PpsbAI::luxAB
(pAM1501, NS1, SpR);
PpsbAI::luxCDE
(pAM1619, NS2, KmR)
This study
AMC1484 kaiA MuCm
insertion
- PkaiBC::luxAB
(pAM1887, NS1, SpR);
PpsbAI::luxCDE
(pAM1619, NS2, KmR)
This study
AMC1485 Wild-type PkaiA::kaiA
(pAM2246, NS1)
PkaiBC::luc
(pAM2105, NS2, CmR)
This study
AMC1486 kaiA ΩKm
insertion
PkaiA::kaiA
(pAM2246, NS1)
PkaiBC::luc
(pAM2105, NS2, CmR)
This study
AMC1487 Wild-type PkaiA::kaiA281
(pAM3434, NS1)
PkaiBC::luc
(pAM2105, NS2, CmR)
This study
AMC1488 kaiA ΩKm
insertion
PkaiA::kaiA281
(pAM3434, NS1)
PkaiBC::luc
(pAM2105, NS2, CmR)
This study
AMC1491 kaiA MuGm
insertion
- PkaiBC::luc
(pAM2105, NS2, CmR)
This study
AMC1492 kaiA ΩKm &
MuGm insertions
- PkaiBC::luc
(pAM2105, NS2, CmR)
This study
AMC1531 kaiA MuGm
insertion
PkaiA::kaiA
(pAM2246, NS1)
PkaiBC::luc
(pAM2105, NS2, CmR)
This study
AMC1532 kaiA MuGm
insertion
PkaiA::kaiA281
(pAM3434, NS1)
PkaiBC::luc
(pAM2105, NS2, CmR)
This study
AMC1533 kaiA ΩKm &
MuGm insertion
PkaiA::kaiA
(pAM2246, NS1)
PkaiBC::luc
(pAM2105, NS2, CmR)
This study
AMC1534 kaiA ΩKm &
MuGm insertion
PkaiA::kaiA281
(pAM3434, NS1)
PkaiBC::luc
(pAM2105, NS2, CmR)
This study
∗Cm, chloramphenicol; Km, kanamycin; Sp, spectinomycin; Gm, gentamycin
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
Table 2. Bacterial plasmids Plasmid Characteristics Antibiotic resistance ∗ Source or reference
pAM2246 PkaiA::kaiA (NSI) SpR, Sm
R Ditty et al. (2005)
pAM3434 PkaiA::kaiA281 (NSI) SpR, Sm
R This study
pAM3515 pHP45ΩGm GmR, Ap
R This study
pAM3582 kaiA::MuCm GmR, Ap
R This study
pAM3613 kaiA::MuGm GmR, Ap
R This study
pAM3633 pET-32a(+)-kaiA ApR Vakonakis et al. (2004)
pAM3630 pET-32a(+)-kaiA281 ApR This study
∗Gm, gentamycin; Km, kanamycin; Sp, spectinomycin; Sm, streptomycin; Ap, ampicillin
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from
Table 3. Circadian clock period of wild-type and mutated kaiA strains PCC 7942
Strains
Original kaiA
locus*
Ectopic kaiA
(NSI)†
KaiB (KaiC)
protein level Period±±±±SEM‡
(h)
N
AMC1020 WT - normal 24.40±0.05 39
AMC1483 281 (MuCm) - ↑ 23.32±0.05 57
AMC1300 WT - normal 25.35±0.04 41
AMC1484 281 (MuCm) - ↑ 24.42±0.03 48
AMC541 WT - normal 25.03±0.02 54
AMC1161 Null (ΩKm) - - AR 20
AMC1485 WT WT ↑ 24.12±0.03 25
AMC1487 WT 281 ↑ 23.81±0.03 26
AMC1486 Null (ΩKm) WT normal 25.02±0.03 28
AMC1488 Null (ΩKm) 281 normal 24.49±0.05 23
AMC1491 281 (MuGm) - ↑ 24.51±0.05 23
AMC1492 Null (ΩKm, MuGm) - - AR 24
AMC1531 281 (MuGm) WT ↑ 24.48±0.10 17
AMC1532 281 (MuGm) 281 ↑ 23.81±0.08 16
AMC1533 Null (ΩKm, MuGm) WT ↑ 25.62±0.04 25
AMC1534 Null (ΩKm, MuGm) 281 ↑ NA§ 16
* WT: PkaiA::kaiA, wide-type allele; 281: PkaiA::kaiA281, the truncated allele; MuCm: the original Mu insertion at
the 3’-end of kaiA coding region; ΩKm: the omega cassette inserted at the 5’-half of kaiA for construction of a null
strain; MuGm: the reconstructed Gm version of the original Mu insertion.
† PkaiA::kaiA or PkaiA::kaiA281 alleles introduced in NS1.
‡ SEM: standard error of the mean
§ NA: not available; unstable traces, phase advanced and low amplitude (see Fig. 3A)
on February 12, 2018 by guest
http://jb.asm.org/
Dow
nloaded from