|
|
|
|
|
|
Journal of Bacteriology, June 2004, p . 3889-3902, Vol . 186, No . 12
Consequences of a Deletion in dspA on Transcript Accumulation in
Synechocystis sp . Strain PCC6803
|
| ABSTRACT |
|---|
A sensor histidine kinase of Synechococcus sp . strain PCC7942,
designated nblS, was previously identified and shown to be critical
for the acclimation of cells to high-light and nutrient limitation
conditions and to influence the expression of a number of light-responsive
genes . The nblS orthologue in Synechocystis sp . strain
PCC6803 is designated dspA (also called hik33) . We have
generated a dspA null mutant and analyzed global gene
expression in both the mutant and wild-type strains under high- and
low-light conditions . The mutant is aberrant for the expression of
many genes encoding proteins critical for photosynthesis, phosphate
and carbon acquisition, and the amelioration of stress conditions .
Furthermore, transcripts from a number of genes normally detected
only during exposure of wild-type cells to high-light conditions
become partially constitutive in the low-light-grown dspA
mutant . Other genes for which transcripts decline upon exposure of
wild-type cells to high light are already lower in the mutant during
growth in low light . These results suggest that DspA may influence
gene expression in both a positive and a negative manner and
that the dspA mutant behaves as if it were experiencing stress
conditions (e.g., high-light exposure) even when maintained at
near-optimal growth conditions for wild-type cells . This is discussed
with respect to the importance of DspA for regulating the responses
of the cell to environmental cues .
| INTRODUCTION |
|---|
Photosynthetic organisms have evolved intricate mechanisms for
sensing and acclimating to environmental change . Parameters such as
light quality, light intensity, and nutrient availability can
modulate both the structure and the function of the photosynthetic
machinery . Physiological and biochemical changes elicited by external
cues include modification of light-harvesting complex (LHC) synthesis
and degradation (6, 12, 27,
29, 38, 45,
65, 75), changes in absorption
and excitation energy transfer properties of LHC (11,
18, 28, 59), and a
modification of reaction center function (59,
80) . Intracellular cues critical for controlling
cellular processes during acclimation may reflect the cell's growth
potential, cellular redox conditions, and/or accumulation of reactive
oxygen species (52, 77) .
Precise control over the fate of absorbed excitation energy is critical for cell viability during exposure of photosynthetic cells to excess excitation, since energized pigment molecules may trigger the production of damaging, reactive oxygen species (2) . Over the short term, cells can dissipate excess absorbed excitation energy as heat by quenching excited pigment molecules in the LHC or by eliciting a state transition in which the LHC of photosystem II (PS II) directs its excitation energy to PS I, where quenching can occur . Over the long term, excess excitation may cause a dramatic reduction in the level of LHC . In cyanobacteria, a reduction in LHC size is reflected in reduced levels of transcripts encoding light-absorbing polypeptides (or phycobiliproteins) of the major LHC (or phycobilisomes) . Several studies have shown that light and nutrient conditions modulate levels of phycobiliprotein transcripts (e.g., from the cpcBA and apcAB genes) (9, 30, 40, 41, 53, 60), enzymes required for pigment biosynthesis (e.g., from hem and chl genes), and subunits of the photosynthetic reaction centers (e.g., from psaA, psaB, psbA, and psbD genes) (36) .
PS II is a primary site of damage to the photosynthetic machinery in high light (HL), causing photoinhibition . The PS II reaction center polypeptides are composed of D1 and D2, and damage to the reaction center reflects a rapid loss of function of D1 . Cellular processes have evolved to facilitate rapid turnover and exchange of photodamaged D1 subunits (10) . In cyanobacteria, the D1 and D2 polypeptides are encoded by two gene families, psbA and psbD, respectively . Synechococcus sp . strain PCC7942 has three distinct psbA genes and two distinct psbD genes . The three psbA genes encode two forms of D1 (7, 25); psbAI encodes D1, form I, which is abundant under low-light (LL) conditions, while psbAII and psbAIII each encode D1, form II, which becomes abundant when cells are shifted from LL to HL . The latter form of D1 is important for maintaining PS II function and the fitness of cells under HL conditions (1, 8, 31, 50, 51) . Synechocystis sp . strain PCC6803 also has three psbA genes, but the expression of only psbAII and psbAIII has been observed; transcripts from both psbAII and psbAIII increase during exposure of the cells to HL (13) . Synechocystis sp . strain PCC6803 also has two psbD genes (psbDI and psbDII) encoding D2 (35, 56) .
Another family of cyanobacterial genes that are highly expressed during HL exposure and nutrient deprivation has been designated hli (HL induced) (15) or scp (small chlorophyll binding protein) . The genes in this family have homology to light-harvesting polypeptides of vascular plants but have one membrane-spanning helix, rather than three (15) . In Synechocystis sp . strain PCC6803, there are four hli genes, hliA, hliB, hliC, and hliD (also called scpB, scpC, scpD, and scpE) (21), as well as an hli sequence fused to the ferrochelatase coding region (21, 34) . All of the Hli polypeptides accumulate following the exposure of cells to HL conditions, although the kinetics of accumulation may be different . These polypeptides also accumulate under conditions of nitrogen or sulfur deprivation and low temperature, suggesting that they are stress-related polypeptides that are synthesized when cells absorb excess excitation energy (34) . A null mutant for all four of the hli genes (hli quadruple mutant) grows well in LL but, unlike wild-type cells, dies upon exposure to HL (34) . The Hli polypeptides may be involved in protecting cells from HL damage and may mechanistically facilitate the dissipation of excess absorbed light energy (33) and/or limit the accumulation of free phototoxic intermediates in chlorophyll biosynthesis (81) .
Recently, we isolated a photosynthetic mutant of Synechococcus sp . strain PCC7942 that exhibited abnormal expression of the light-regulated hli, psbA, and cpcBA genes and that died rapidly following exposure to either HL or nutrient deprivation . The gene altered (by two point mutations) in the mutant encodes a sensor histidine kinase (62, 71, 72), designated NblS (77), that has a PAS domain and may interact with the response regulator NblR (69) . The PAS domain may bind a chromophore and function in monitoring light, oxygen, and/or redox conditions (76, 77) . We were unable to obtain an nblS null mutant of Synechococcus sp . strain PCC7942 (77) . Recently, a mutant of Synechocystis sp . strain PCC6803 that was null for dspA (also called hik33) (73) was isolated, and the mutant phenotype was rescued by the Synechococcus sp . strain PCC7942 nblS gene (42) . Like the nblS mutant, the dspA strain showed aberrations in expression of the psbA and hli genes, although the phenotype of the dspA null mutant was somewhat different than that of the nblS mutant .
Microarray-based methods for high-throughput monitoring of gene expression have been developed over the last decade (14, 66, 82) . Such methods are powerfully used for analyses of gene expression in organisms for which there is a complete genome sequence, such as Synechocystis sp . strain PCC6803 (46) . Recent microarray studies have exploited Synechocystis sp . strain PCC6803 for identifying genes whose activities are modulated by light levels and UV irradiation (22, 36, 43) . To further define the role of NblS/DspA/Hik33 (3, 54, 73) in the responses of cells to environmental change, we used microarrays to compare global transcript levels in wild-type Synechocystis sp . strain PCC6803 and the dspA mutant both in LL and following exposure of cells to HL . The results suggest that DspA has a far-reaching effect on the levels of numerous transcripts, including those encoding polypeptides directly or indirectly associated with photosynthesis and other metabolic processes, and that these effects are apparent under both HL and LL conditions .
| MATERIALS AND METHODS |
|---|
Strains and growth conditions. The nonmotile strain of
Synechocystis sp . strain PCC6803 and the dspA mutant (42)
were cultured in BG-11 medium with or without supplementation with 10
mM glucose . Cells were maintained in LL (30 µmol of photon m–2
s–1) and bubbled with 3% CO2 in air . For HL
treatment, cells were grown according to a method described
previously (34) . Briefly, when cultures attained
mid-logarithmic growth (optical density at 750 nanometers of
0.6),
they were diluted to an optical density at 750 nanometers of
0.3
and either placed back in LL or illuminated with 500 µmol of photon m–2
s–1 of white light for various periods of time . The cells
were then cooled in liquid nitrogen and collected by centrifugation
(5,000 x g for 5 min), and
the cell pellets were stored at –80°C until they were used for RNA
isolation .
Amplification of gene fragments and array construction. The
whole set of PCR-amplified gene-specific fragments used to construct
the microarray was generated as previously described (63);
all primers used contained an additional universal sequence (forward
primer, 5'-CCAGGGGAGCTCTTCCATG-3'; reverse primer,
5'-GCACGGGTGCTCTTCCACC-3') that allowed for the reamplification of
the fragments with a single primer pair . The PCRs were performed in
384-well microtiter plates with 0.4 U of Pfx polymerase (Invitrogen,
Carlsbad, Calif.), 10 to 30 ng of the original amplified DNA
fragment, 5 µmol of oligonucleotides, 50 µmol of deoxynucleoside
triphosphates, and 1.5 mM MgCl2 in a 20-µl reaction
volume . For PCR amplification of specific gene fragments from
purified Synechocystis sp . strain PCC6803 genomic DNA (74),
reactions were performed in 96-well microtiter plates using 2 U
of Taq polymerase (MBI Fermentas, Amherst, N.Y.), 100 ng of
genomic DNA, 5 µmol of oligonucleotides, 50 µmol of deoxynucleoside
triphosphates, and 1.5 mM MgCl2 in a 100-µl reaction
volume . Reactions were cycled 25 times at 94°C for 2 min, 92°C for 30
s, 55°C for 30 s, 72°C for 2 min, and, after the final cycle, 72°C
for 5 min . Approximately 5% of the PCR products contained multiple
DNA fragments; these fragments were reamplified either at different
annealing temperatures or from electrophoretically resolved fragments
that were purified from agarose gels . There were still 67 genes for
which products did not amplify well; 70-mer oligonucleotides specific
for these genes were included on the array . The PCR products,
resolved on 1.2% agarose gels and sized with respect to molecular
mass markers prepared by Life Technology (Rockville, Md.), were
purified by MultiScreen PCR 384-well filter plates (Millipore,
Bedford, Mass.), and the nucleic acid concentration was measured with
the PicoGreen double-stranded DNA quantification reagents (Molecular
Probes, Eugene, Ore.) . Purified PCR products were resuspended
in 3x SSC (1x
SSC is 0.15 M NaCl plus 0.015 M sodium citrate) to a final
concentration of 250 to 300 ng/µl . Eight microliters of each PCR
product was transferred from 384-well microtiter plates to the
printing plates, also in the 384-well microtiter format (Genetix
Limited, Hampshire, United Kingdom) . The PCR products were arrayed
onto CMT-GAPS slides coated with
amino propyl silane (Corning Incorporated, New York, N.Y.) by using
an Omnigrid Microarrayer (GeneMachines, San Carlos, Calif.)
with ChipMaker 2 pins (TeleChem International, Sunnyvale, Calif.) .
Four complete cDNA sets, each consisting of 3,707 array elements,
were printed onto each slide . Printed slides were allowed to dry for
12 h at room temperature (25°C) at <50% relative humidity and then
cross-linked by UV irradiation in the Stratalinker 1800 (Stratagene,
La Jolla, Calif.) at a total power of 300 mJ . Just prior to
prehybridization, arrays were baked at 65°C for 10 min . Following
prehybridization for 1 h at 50°C in 3x
SSC-0.1% sodium dodecyl sulfate (SDS)-0.1 mg of bovine serum
albumin/ml, the arrays were rinsed by immersion in double-distilled H2O,
transferred to isopropanol for 1 min, dried by centrifugation for 5
min in a SpeedVac Plus model SC210A (Savant, Holbrook, N.Y.), and
then stored in a desiccator at room temperature until they were used .
Nucleic acid isolation. RNA was isolated from 50-ml cultures
of wild-type cells and the dspA mutant; cells were collected
by centrifugation (5,000 x g
for 5 min), and cell pellets were resuspended in 500 µl of NAES
extraction buffer (50 mM sodium acetate [pH 5.1], 10 mM EDTA, 1%
SDS) . An equal volume of acidic phenol (equilibrated with NAES) was
added to the suspension (approximately 1.5 ml) along with 100 mg of
glass beads (0.1-µm average diameter; Sigma, St . Louis, Mo.) prior to
agitation (two times for 30 s each) at the maximum speed (5,000
rpm) in a Mini-Bead Beater (Biospec Products, Bartlesville, Okla.) .
The beads and unbroken cells were pelleted by a brief centrifugation
(5 to 10 s at full speed in an Eppendorf centrifuge 5415C), and the
supernatant was subjected to two phenol-chloroform (1:1) and one
chloroform extractions . Two volumes of ethanol were added to the
aqueous phase of the final chloroform extraction, and the solution
was incubated for 2 h at –20°C . Precipitated RNA was collected
by centrifugation (10 min at full speed in an Eppendorf Microfuge),
washed with 70% ethanol, resuspended in Tris-EDTA, and treated
with 25 U of RNase-free DNase I (Amersham, Piscataway, N.J.) at room
temperature for 30 min . The DNase I was removed by two
phenol-chloroform extractions, and the RNA was precipitated with 2
volumes of ethanol .
Microarray hybridization and data analysis. Hybridization to the microarray was performed according to the method described by Schena et al . (66), with minor modifications . First-strand cDNA was synthesized by incubating 25 µg of total RNA with mixed gene-specific reverse primers and 30 U of Superscript II RNase H-minus reverse transcriptase (Invitrogen) . The reaction was performed at 42°C for 1 h in a volume of 40 µl . After digesting total RNA with RNase H, second-strand synthesis was accomplished by using the Klenow fragment of DNA polymerase (USB, Cleveland, Ohio) and random primers; Cy3- or Cy5-dUTP fluorescent nucleotides were incorporated into the second strand (Amersham) over a labeling period of 3 h at 37°C . Labeled DNA fragments were purified by using the QIAquick spin column (QIAGEN, Valencia, Calif.), and the cDNA derived from 6.25 µg of total RNA was resuspended in hybridization buffer (3x SSC, 0.3% SDS, 0.17 µg of tRNA/µl) to a final volume of 30 µl . The fluorescent cDNA was denatured at 95°C for 90 s, and its application to the array was done by capillary delivery under a glass lifter slip (Eric Scientific, Portsmouth, N.H.) . Three spots each of 20 µl of 3x SSC were positioned at the edges of the slides to maintain humidity during hybridization . Slides were sealed in custom-built hybridization chambers, and hybridizations were done at 65°C for 16 h . Following hybridization, the arrays were washed for 2 min at room temperature in low-stringency wash buffer (2x SSC, 0.03% SDS), 2 min in moderate-stringency wash buffer (1x SSC), and 2 min in high-stringency wash solution (0.05x SSC); dried in a Speedvac Plus (Savant); and scanned with a GenePix Laser Scanner 4000 (Axon, Foster city, Calif.) . The TIFF images generated by the GenePix Laser Scanner 4000 were imported into the GenePix Pro 3.0 program (Axon), and spot intensities from scanned slides were quantified . Grids were predefined and manually adjusted to ensure optimal spot recognition . Spot signals that were distorted by dust or local high backgrounds were flagged and not included in the analysis . After determining the relative ratios of the fluorescent signals from the two channels (532 and 635 nm), the data files (.gpr files) were imported into the Stanford Microarray Database (http://genome-www.stanford.edu/microarray/) (26, 70) and normalized by using an internal control, rps1b (for details, see the Stanford Microarray Database website) . The primary data sets may be viewed and downloaded (see above); the experiment identification numbers are 38980, 38981, 38985 to 38988, 39071, 39080, 39085, 39086, 39563, 39578, and 39583 to 39590 .
Only those spots with a minimum normalized net intensity for the
median (channel 1 net median intensity or normalized channel 2 net
median intensities) of >350 for both wild-type cells and the dspA
mutant, in at least one of the two channels, were included in the
analyses . Log (base 2) channel 1/channel 2 ratios after normalization
were calculated . The significance of the differences in gene
expression observed (at P = 0.05) was statistically evaluated
by t tests using a df of 7 with an
of 2.8, representing P = 0.05 . The degree of freedom (df)
value was calculated from the 8 samples used for the data analysis;
for each data point, two slides were analyzed, with four duplicates
for each array element on each slide . Two independently isolated RNA
samples were used for generating the cDNA that was hybridized to each
of the slides .
Microarray experimental design. The effect of glucose on gene expression in wild-type cells and the HL effects on both the wild-type and the dspA mutant of the Synechocystis sp . strain PCC6803 were examined . For measuring long-term acclimation of wild-type cells to glucose, total RNA was extracted from cells cultured in BG-11 medium or BG-11 medium supplemented with 10 mM glucose . The cells were grown for 2 weeks in LL but were diluted every 2 days to maintain their logarithmic growth . Total RNA isolated from wild-type cells grown in BG-11 medium in the absence of glucose was reverse transcribed, and the second-strand cDNA was labeled with Cy3-dUTP . Total RNA isolated from the cells grown in BG-11 medium supplemented with glucose was used for labeling cDNA with Cy5-dUTP . The Cy3- and Cy5-labeled cDNA samples were mixed and applied to a microarray slide . A "dye-swap" Cy-dye-dUTP labeling was performed with RNA isolated from a duplicate, independent experiment . A similar procedure was used for analyzing transcripts in wild-type cells and the dspA mutant grown in glucose-supplemented medium following transfer from LL (zero time point reference) to HL (for 30 min, 1 h, 3 h, and 6 h) . We also examined differential transcript accumulation in the dspA mutant in LL compared to that of wild-type cells grown under the same conditions . Total RNA from the LL-grown dspA mutant was reverse transcribed, and the second-strand cDNA was labeled with Cy5-dUTP and mixed with Cy3-dUTP-labeled cDNA generated from LL-grown wild-type cells and applied to a microarray slide . The experiment was duplicated, but the dyes used to label the cDNAs were swapped; the cDNA generated from LL-grown dspA mutant was labeled with Cy3-dUTP, while the cDNA from LL-grown wild-type cells was labeled with Cy5-dUTP . The dspA mutant grew extremely slowly and was difficult to maintain in the absence of glucose .
Real-time quantitative RT-PCR. To confirm the validity of the results generated from microarray analyses, we performed real-time reverse transcription (RT)-PCR for 8 specific transcripts and compared the results to the mean of the quantified array measurements for those same transcripts . To remove trace amounts of genomic DNA contamination from RNA samples used for the real-time RT-PCR, RNA (100 ng/µl) was treated with 2 U of RNase-free DNase at 37°C for 30 min (Ambion, Austin, Tex.), and the DNase was removed by DNase removal reagent (Ambion) . Levels of specific transcripts in total mRNA isolated from HL-treated cells were quantified by real-time RT-PCR with the LightCycler (Roche Applied Science, Indianapolis, Ind.) . Amplifications were performed by incubating 200 ng of DNA-free RNA with specific primer pairs, as given in Table 1, and LightCycler RNA Master SYBR Green I according to the protocols of the manufacturer (Roche Applied Science) . Amplifications performed according to the protocols are as follows: 1 cycle of RT reaction at 61°C for 20 min, 1 cycle at 95°C for 2 min, and 45 cycles at 95°C for 5 s, 55°C for 5 s, and 72°C for 13 s . A "no-RT" PCR control reaction was performed by incubating an equal amount of DNA-free RNA samples with a mixture of specific primer pairs and LightCycler DNA Master SYBR Green I at a final volume of 20 µl (Roche Applied Science) . Amplification was done for 1 cycle at 95°C for 30 s followed by 45 cycles of 95°C for 5 s, 55°C for 5 s, and 72°C for 10 s .
|
A number of standards were developed as controls for these experiments .
The rps1b (slr1984) and rnpB (slr0249)
transcripts were used as external standards since the levels of these
transcripts did not change in either wild-type cells or the dspA
mutant following HL exposure (under all conditions that we used) . PCR
efficiency differences between target and reference transcripts
were corrected by generating fit coefficients . A dilution series was
used for constructing standard curves that enabled us to determine
the fit coefficients for PCR amplification efficiency of target
relative to the reference (rps1b and rnpB) transcripts .
Quantification of the relative change (n-fold) in mRNA levels
was calculated as described by Technical Note no . LC13/2001 (Roche
Applied Science, 2001) . Each quantitative RT-PCR experiment was
performed with at least two replicates .
| RESULTS |
|---|
(i) Effect of glucose on expression in wild-type cells and the dspA
mutant. In LL, the dspA mutant exhibited a markedly reduced
growth rate (doubling time of
40
h) under photoautotrophic conditions relative to wild-type
Synechocystis sp . strain PCC6803 and was difficult to maintain .
This slow growth rate reflected very low light-dependent O2
evolution (42), demonstrating that a strain lacking DspA
is strongly affected in both growth and photosynthetic performance .
In contrast, in the presence of glucose, the mutant grew, with
a doubling time of 8 h, compared to a 6-h doubling time for wild-type
cells (42) . We examined how glucose influenced global
transcript accumulation in wild-type Synechocystis sp . strain
PCC6803 . Cells were grown long term (2 weeks with the dilution
of cultures to maintain the cells in logarithmic growth phase) in LL
in either BG-11 or BG-11 supplemented with 10 mM glucose . Total RNA
was purified (5) and reverse transcribed, second-strand
cDNAs were labeled with Cy3- or Cy5-dUTP, and equal amounts of
the two labeled cDNA samples were hybridized to the arrays (see
Materials and Methods) . The use of second-strand synthesis for probe
preparation, compared to using just first-strand synthesis, increased
the array signals generated without significantly changing the
relative intensities of the signals .
Of the 3,164 distinct Synechocystis sp . strain PCC6803 genes
represented on the array,
95%
showed no statistically significant difference in transcript levels
between wild-type cells grown photoautotrophically and those grown
mixotrophically (supplemented with 10 mM glucose) . Statistically
significant changes in the levels of transcripts for 5% of the genes
were relatively small; all changes were less than twofold (data not
shown) . These results demonstrate that during long-term growth, the
pattern of transcript accumulation in wild-type cells is not
dramatically influenced by the inclusion of glucose in the growth
medium . Therefore, we compared the effects of HL on gene expression
in wild-type cells and the dspA mutant during growth in
glucose-supplemented medium .
(ii) Gene expression in wild-type cells and the dspA mutant. Expression from 3,164 Synechocystis sp . strain PCC6803 genes was measured for wild-type cells and the dspA mutant grown in LL and then transferred to HL for 0.5, 1, 3, and 6 h . A comparison of transcript abundance in wild-type cells and the dspA mutant grown in LL demonstrated that the levels of 84 transcripts were altered by more than twofold in the mutant relative to wild-type cells (data not shown) . Nearly all of these transcripts showed increased abundance in the dspA mutant relative to wild-type cells during LL growth, with the exceptions of chlL, chlN, hofG, cpcC, psaA, psaB, cpcG, sll0177, and slr0147, for which the transcripts were lower in the mutant strain .
We identified a set of 143 genes for which the transcript abundance changed by at least threefold in either wild-type cells or the dspA mutant following the transfer of cells to HL (for at least one of the time points following transfer) or that were already elevated in the dspA mutant compared to wild-type cells during LL growth . A hierarchical clustering analysis was used to group the genes on the basis of similarities in their expression patterns and to allow for easy viewing of the changes that we observed; six different groups were established, as shown in Fig . 1 . For each gene in the different expression categories, the change in transcript levels over the 6-h time course following the transfer of cells to HL (for both wild-type and dspA mutant cells), the difference in the initial expression levels for wild-type cells and the dspA mutant under LL conditions (0 h), and the statistical significance of the data are given in Table 2 . Table 3 shows trends in expression within specific categories of genes critical for photosynthesis and acclimation of cells to phosphorus deprivation . In this table, all changes of twofold or greater are shown .
|
|
|
Group 1: transcripts in wild-type cells increased, while dspA
transcripts declined. Seven genes were placed in group 1 . When
wild-type cells were exposed to HL, the level for all of the seven
genes increased to a significant extent, while they declined by 30 to
70% in the dspA mutant . The genes in this group include two
genes related to cell division, ftsH (or slr0228) and
slr1604 . The ftsH gene encodes a metalloprotease
involved in the replacement of damaged D1 polypeptides in the
reaction center of PS II 46a, 64a, 70a) . The other genes in this
group include psbD2 (slr0927), which encodes a second
reaction center polypeptide; pstS (sll0680), which
encodes a phosphate binding protein; pstC (sll0681), which
encodes a subunit of the phosphate permease; and two genes encoding
proteins of unknown function (Table 2) . Interestingly,
transcript levels of genes in this category were 2.6- to 17-fold
higher in the mutant than in wild-type cells during LL growth; the
pstS and the pstC transcripts were 17- and 10-fold more
abundant, respectively, in the mutant than in wild-type cells during
LL growth .
Group 2: transcripts that changed slightly in wild-type cells and declined in the dspA mutant. Surprisingly, most of the phosphate starvation-induced genes (pho), extensively studied in E . coli (78), were in group 2; this group includes pstA (slr1249), pstB (sll0684, sll0683), and pstS (slr1247) . Some transcripts for pho regulon polypeptides exhibited a transient increase in HL, as is clearly observed for pstC and pstS (Tables 2 and 3) . A similar increase in pstS mRNA in HL was observed previously (4) . Interestingly, even in LL, the dspA strain showed very high levels (relative to wild-type cells) of pstS (both sll0680 and slr1247), pstC (sll0681 and slr1248, integral membrane protein), and pstB (sll0683 and slr1250, nucleotide binding protein) transcripts, as well as of transcripts encoding an alkaline phosphatase (sll0654) and the extracellular NucH nuclease (sll0656) (Table 3) . Although nucleases have been associated with the pho regulon, it is not known whether the nucH transcript increases in response to phosphorus limitation . However, it is likely to be part of the pho regulon since it is immediately downstream of the alkaline phosphatase and the two genes may be part of the same operon . These data demonstrate the apparent derepression of several genes associated with phosphorus starvation in the dspA mutant in LL . There was no further increase in the levels of pho transcripts when the dspA mutant was exposed to HL, and the abundance of some of the transcripts declined significantly (Table 3) .
Interestingly, while the transcript encoding phosphoglycolate phosphatase (CbbZp, sll1349) was not elevated significantly in wild-type cells during HL treatment, it was fivefold higher in the mutant relative to wild-type cells under LL growth conditions . This enzyme is part of the photorespiratory pathway, which can serve as an energy sink that prevents an overreduction of the photosynthetic electron transport chain (and consequent photoinhibition) and may be especially valuable under stress conditions that lead to reduced CO2 fixation (79) . Photorespiration also provides metabolites for the production of glutathione, a tripeptide that protects cells from oxidative damage . Other genes in this category included htrA (slr1204; 4.7-fold), encoding a serine protease; rfbP (sll1535), encoding a putative sugar transferase; and six genes encoding proteins of unknown function .
Group 3: levels of transcript in both wild-type cells and the dspA
mutant strongly declined. Genes associated with photosynthetic
function, especially those encoding phycobilisomes and PS I
polypeptides, are in group 3 . The levels of most cpc
transcripts showed a marked decrease in both wild-type and dspA
mutant cells following the imposition of HL conditions (Tables
2 and 3) . The cpcE and cpcF
transcripts were exceptions to this trend . These transcripts code for
subunits of a lyase involved in chromophorylation of the
-PC
subunit (16, 17) .
Interestingly, most cpc transcripts (cpcA, cpcB, cpcC,
cpcD, and cpcG) were already 25 to 50% lower in the mutant
than in wild-type cells during LL growth (Table 3) .
In contrast, most apc transcripts were equally abundant in
LL-grown mutant and wild-type cells . There was a pronounced transient
change in apc transcript levels following exposure of
wild-type cells to HL, but by 6 h in HL, the transcript levels
approached those observed for LL-grown cells .
The transcripts for most PS I polypeptides (including psaA, psaB, psaF, and psaL) were already 30 to 60% lower in the dspA mutant than in wild-type cells in LL . The most pronounced difference between the mutant and wild-type cells was for the psaA/psaB transcripts, which were 60% lower in LL-grown mutant cells (Tables 2 and 3) . Furthermore, the levels of many transcripts declined in wild-type cells following exposure to HL and stayed low (psaA, psaB, psaC, psaD, psaF, psaJ, and psaL), while others didn't decline nearly as much or showed a transient decline (psaE, psaI, and psaK) . For a number of PS I genes, the decline in the dspA mutant was not quite as severe as is it was in wild-type cells (e.g., psaD, psaF, and psaJ) (Table 3, compare the wild type and the dspA mutant) .
Transcripts encoding polypeptides associated with chlorophyll biosynthesis are also in this category . Transcripts encoding two of the subunits of the light-independent prochlorophyllide reductase (chlL [slr0749] and chlN [slr0750]) were already reduced in the dspA mutant relative to wild-type cells (30 to 40%) in LL . While chlL and chlN transcripts declined markedly following the exposure of wild-type cells to HL (Table 2), they did not decline to nearly the same extent in the mutant (they were already low in the mutant strain in LL) . The other genes in this category with characterized functions were nrtB (sll1451), encoding a nitrate/nitrite permease; crhR (slr0083), encoding a RNA helicase; and ho1 (sll1184), encoding a heme oxygenase .
Group 4: transcript that markedly declined in wild-type cells but increased in dspA. A small number of genes were clustered in group 4, which included genes encoding the pilin polypeptide (hofG or pilA [sll1694 and sll1695]), psaD, rbpA (sll0517), and lexA (sll1626) (Table 2) .
Group 5: transcripts that declined slightly in wild-type cells but increased in dspA. The transcripts for most genes encoding constituents of the carbon-concentrating mechanism (CCM) (47) were slightly higher in the dspA mutant than in wild-type cells in LL conditions . Furthermore, the levels of most of these transcripts did not change much when wild-type cells were transferred to HL, although the ccmM and ccmK transcripts may have declined slightly . Interestingly, there was a relatively large transient increase in the levels of transcripts (four- to sixfold) from a number of the CCM genes (ccmK [sll1028 and sll0129], ccmL, ccmM, and ccmN) (Table 2) following exposure of the dspA mutant to HL, suggesting that the mutant cells were starved for CO2 immediately following the transfer to HL . There was essentially no difference in the levels of rbcL and rbcS transcripts (and in transcripts encoding components of the reductive pentose phosphate pathway; data not shown) between wild-type cells and the dspA mutant grown in LL, and HL did not significantly affect the abundance of these transcripts in wild-type cells, although there may have been a small increase in rbcL and rbcS transcript levels following the exposure of dspA to HL . Furthermore, small but significant increases in the prk transcript (encoding phosphoribulokinase) were observed in both wild-type and mutant cells following HL exposure .
Group 6: transcripts that increased in both wild-type cells and dspA. There were a total of 69 genes whose transcript levels were increased significantly in both wild-type cells and the dspA mutant following the transfer from LL to HL . These genes included those with functions associated with photosynthesis (PS II, NADH dehydrogenase, and ATP synthase), HL and heat shock acclimation processes, and ribosome biosynthesis . Importantly, the transcripts from a number of these genes were already higher in the LL-grown dspA mutant relative to LL-grown wild-type cells . As presented in Tables 2 and 3, the level of the psbA transcripts (we were not able to distinguish between psbAII and psbAIII transcripts because they are nearly identical) was five- to sixfold higher in the mutant than in the wild-type strain grown under LL conditions . Similarly, the levels of psbDI and psbDII transcripts were 2.6- to 3.4-fold more in the dspA mutant than in wild-type cells grown in LL (Tables 2 and 3) . Similar to the psbA genes, the psbDI and psbDII genes show strong sequence similarity, and substantial cross-hybridization of the cDNA probes to the psbD array elements almost surely occurred .
Upon the transfer of wild-type cells from LL to HL, the levels of transcripts for some PS II polypeptides showed a transient increase and then declined to near-LL levels (psbDI and psbDII) (Table 3), while others did not change very much . However, the psbA transcript levels went up sharply (approximately 5-fold) but declined to 2.5-fold of the LL level following a 6-h exposure to HL . Transient features of psbA transcript accumulation in cyanobacteria following HL exposure were previously described (24, 50) . The changes in transcript levels for the psbA genes in the dspA mutant were not as pronounced as those in wild-type cells, and while there was some increase in psbA mRNA at early times following HL imposition (between 1.5- and 2-fold), the change was only transient, and the final level after 6 h in HL was similar to the level in LL (Table 3) (although, as already mentioned, the initial levels were much higher in the dspA mutant than in wild-type cells in LL) .
Transcripts for all hli genes were also higher in the dspA mutant than in wild-type cells in LL, which had been previously observed from RNA blot hybridizations (42); the transcript for hliC showed the highest increase (approximately sevenfold) relative to wild-type cells (Table 3) . Furthermore, there was a large transient increase in the levels of hliB and hliC transcripts in wild-type cells following 1 h of HL exposure (7- to 10-fold), with a subsequent decline; but the level still remained well above the LL level . The dspA mutant also exhibited elevated levels of hli transcripts following the exposure of cells to HL, although the relative increases were not as great as those observed for wild-type cells (Table 3) .
Transcripts from many of the genes encoding subunits of the pyridine nucleotide dehydrogenase complex were elevated by 1.4- to 2.0-fold in the dspA mutant relative to wild-type cells grown in LL (Table 2) . Furthermore, for most ndh genes, there was either no significant change in transcript levels or a transient change immediately following exposure of cells to HL (data not shown) . The ndhD2 (slr1291) transcript exhibited a relatively large transient change in wild-type cells at 30 min (3.6-fold) and 1 h (3.8-fold) after HL exposure (the transient change was greater in the mutant than in wild-type cells), but the level dropped to 1.78-fold the LL value after 6 h in HL (Table 2) . The NdhD2 protein has been associated with the transfer of electrons between NADPH and the plastoquinone pool (61), which may be important for controlling cyclic electron flow and redox poise under HL conditions .
Transcripts encoding subunits of the ATP synthase and cytochrome b6f complexes were generally slightly higher in the dspA mutant than in the wild-type strain in LL (data not shown) . However, there were only small, transient effects of HL on levels of transcripts encoding polypeptides of these complexes, with the most significant influence on atpD, atpG, and petF transcripts, which exhibited approximately twofold transient increases .
There were pronounced changes in transcript levels for a number of stress-related genes following the exposure of both mutant and wild-type cells to HL; the transcripts for many of these increased in both mutant and wild-type cells exposed to HL (Table 2, group 6) . The level of groEL1 (slr2076), groEL2 (sll0416), and groES (slr2075) transcripts were increased in both wild-type cells and the dspA mutant by six- to eightfold following HL imposition and were present at only slightly higher levels in the mutant strain during LL growth . The hsp17 mRNA showed a twofold increase following HL exposure but was already high in the dspA mutant in LL (4.5-fold dspA/wild-type cell mRNA) . One of the transcripts encoding glutathione peroxidase (slr1992) was also already elevated by about threefold in LL in the dspA mutant relative to the wild type . The mRNA for the heat shock protein HtpG (sll0430) increased three- to fourfold in both wild-type and mutant cells following HL exposure but was only slightly more abundant in the dspA mutant in LL . Most transcripts for the other hsp genes (including the dnak, dnaJ, and clp genes) were not dramatically affected under the conditions used for these experiments (data not shown) .
A number of other transcripts increased in both wild-type cells and the dspA mutant following exposure to HL (Table 2, group 6); some of these transcripts exhibited large transient increases . These included transcripts encoding a glucanase (slr0897; 12-fold), phosphoglycerate kinase; phosphoribulokinase (prk [slr0394]; 9-fold), which is involved in both photosynthetic and respiratory metabolism; FabG (sll0330; 5.5-fold), which is involved in lipid metabolism; CheY (slr1982; 4.8-fold), a chemosensory protein; HlyD (sll1181; 6-fold), which is involved in secretory processes; and a number of hypothetical proteins encoded by genes of unknown function . Analyzing the activities and functions of the genes encoding hypothetical proteins are likely to unveil novel aspects of acclimation processes .
(iii) Analysis of transcript levels by real-time RT-PCR. We have used real-time RT-PCR to examine transcript levels in wild-type cells and the dspA mutant following the imposition of HL to help evaluate the microarray data . These analyses were performed for some genes whose transcripts increased and others whose transcripts declined following HL exposure . The results of these analyses are presented in Table 4 . Similar to the microarray results, real-time RT-PCR showed that HL stimulated an increase in the levels of groEL-1 and groEL-2 transcripts; the highest levels in wild-type cells were observed 1 h after the imposition of HL . Similar trends were noted for the dspA mutant, although peak transcript levels were at 30 min (for both the microarray and real-time RT PCR analysis) . Furthermore, similar to the groEL results, both microarray and real-time RT-PCR results showed that the psbAIII and pstS transcripts in wild-type cells accumulated to maximal levels after 1 h in HL, although the increase measured by real-time RT-PCR was two to three times that determined by microarray analysis . In contrast, the levels of these transcripts did not increase much in the dspA strain following the imposition of HL; this result probably reflects the fact that they are already high in the mutant strain prior to HL exposure . Finally, the cpcB and psaA transcripts generally declined in both wild-type and mutant cells (as measured by both methods), and the ccmK transcript declined in wild-type cells but showed a strong transient increase in the dspA mutant . Overall, the comparisons of real-time RT-PCR and microarray results were in excellent agreement with respect to the direction of change in transcript abundance for any given gene, while the extent of change determined by these two methods was often significantly different (note that the standard deviations for the real-time RT-PCR are relatively high) .
|
| DISCUSSION |
|---|
Analysis of wild-type cells. Cells grow in environments that
may rapidly fluctuate with respect to light, temperature, and
nutrient availability . Changing environmental conditions will cause
changes in the flux of electrons through the photosynthetic electron
transport system and may also cause an imbalance in electron flow
between the two photosystems, which either directly or indirectly
stimulates a change in gene activity and the activities of specific
protein complexes .
A number of groups have begun to examine the effects of light on
levels of mRNAs in Synechocystis sp . strain PCC6803 (36,
43), as well as in other photosynthetic organisms such
as Chlamydomonas reinhardtii (44) and
Arabidopsis thaliana (49, 64), using
both cDNA and oligonucleotide-based arrays . This "semiquantitative"
technology has been validated with RNA blot hybridizations and
real-time RT-PCR, as presented in this paper (and the companion paper
[42]) . We have been especially interested in how the
photosynthetic machinery changes in response to changing light
conditions . In cyanobacteria, the PS II/PS I ratio generally
increases upon a shift from LL to HL (37,
48, 57) . It was suggested that diminished
PS I content would be expected to lower the susceptibility of
cells to HL damage, particularly during prolonged exposure . Many of
the results obtained with wild-type cells in this study are congruent
with this idea as well as with results obtained in other studies with
Synechocystis sp . strain PCC6803, in which levels of various
transcripts were analyzed following a transfer of cells from LL to
HL . For example, both we an others have demonstrated that levels of
transcripts encoding psbAII/psbAIII and psbD genes
increase following the exposure of Synechocystis sp . strain
PCC6803 to HL (13, 36, 42) .
Increased transcript production may reflect the need for more rapid
synthesis of these proteins, especially D1, which is known to rapidly
turn over under HL conditions . Levels of the cpc, psaA,
and psaB transcripts declined, and levels of apc
transcripts remained relatively constant following the transfer of
cultures to HL conditions . This result suggests that both
light-harvesting capacity and PS I activity are declining under the
light conditions used in this experiment, which agrees with
physiological and biochemical experiments (19,
20) as well as with molecular experiments (58)
performed by others . Transcripts from other genes such as those
belonging to the pho regulon, and especially those involved in
phosphate transport, were shown to increase following the exposure of
cells to HL . This was first observed in studies in which differential
display was used to identify genes responsive to HL conditions (4) .
Differences in values and precise sets of genes modulated by HL
conditions are likely to reflect differences in conditions (and
strains) used for the experiments . For example, such differences are
observed when our data are compared with that of Hihara et al . (36),
with respect to some of the ndh genes (ndhF3 and
ndhD3), some of the ccm genes (ccmM and ccmK),
and the rbcS and rbcL genes . The experiments presented in
this paper result from a transfer of cells from 30 to 500 µmol
of photon m–2 s–1, while Hihara et al . transferred
cells from 50 to 200 µmol of photon m–2 s–1 (36) .
It should be noted that even the HL levels used in the experiments
presented in this study are not detrimental to cell growth (the cells
maintain an optimal doubling time of
6
h), which suggests that modulating components and activities of
the photosynthetic apparatus (and other metabolic processes in
the cell) compensates for the relatively high-intensity light
conditions, limiting photodamage and its potential deleterious
effects on cell growth . Higher light levels might result in more
extreme changes in transcript levels and the activation of additional
genes encoding chaperones and oxidative stress-related polypeptides .
Chaperones and heat shock proteins are also affected by light conditions . Heat shock proteins have been shown to bind unfolded polypeptides and to release them in an ATP-dependent reaction under stress condition such as those of high temperature and HL (32) . These proteins may assist in the proper folding of nascent polypeptide chains as they emerge from ribosomes, participate in transmembrane protein transport, prevent aggregation of denatured proteins, and disassemble multiprotein complexes (32, 39) . There are a number of genes in Synechocystis sp . strain PCC6803 that encode proteins homologous to heat shock proteins from other organisms (23, 46) . Many of these genes were shown to be induced during both heat shock and HL exposure (36, 52a); these genes include clpB, htpG, dnaK (sll0170), groEL-1, groEL-2, groES, and hsp17 . In our experiments, the genes for the chaperones groEL-1, groEL-2, and groES showed strong induction under HL stress; the transcripts for many of the other hsp genes did not increase to very high levels following the exposure of cells to HL . Although the actual function of GroEL or Hsp60 in Synechocystis sp . strain PCC6803 is not clear, recent studies have shown that Hsp70 of C . reinhardtii helps protect cells from photoinhibition; the overexpression of Hsp70 reduced photoinactivation of PS II and enhanced recovery from photoinhibition, while reduced expression caused the opposite effect (67) . The role of the Hsp proteins in protection against photoinhibition had already been proposed based on the finding that the exposure of C . reinhardtii to heat shock resulted in enhanced resistance to light stress (68) .
Analysis of mutant cells. We recently identified the nblS gene (nonbleaching mutant sensor), which is important for the acclimation of cyanobacteria to both HL and nutrient limitation conditions (77) . This gene encodes a sensor molecule predicted to be integral to the membranes and contains a PAS domain that might bind a redox active molecule such as a flavin (this still needs to be clearly established) . The homolog to nblS in Synechocystis sp . strain PCC6803, which is dspA or hik33, is important for Synechocystis sp . strain PCC6803 to acclimate to HL, nutrient limitation, cold, and osmotic conditions (42, 55, 58, 73, 77) . NblS and DspA polypeptides may be global regulators that tune cellular metabolism to the growth potential (77) .
In our initial studies of a dspA null mutant, we demonstrated abnormal regulation of several genes that respond to light intensity (42), and it was therefore important to understand how wild-type Synechocystis sp . strain PCC6803 responded to HL and the relationship between the HL responses observed in the mutant and wild-type strains . In this study, we have demonstrated that numerous transcripts were markedly different in abundance in wild-type Synechocystis sp . strain PCC6803 and the dspA mutant, even under our LL conditions . Importantly, many transcripts that accumulated in wild-type cells in HL were already high in the dspA strain in LL . Furthermore, a number of transcripts that declined in wild-type cells following exposure to HL were already low in LL-grown dspA . Such changes include increases in levels of transcripts encoding PsbAII/PsbAIII, PsbD, the Hli polypeptides, the HtpG and to some extent the GroEL/GroES chaperones and decreases in transcripts encoding the PC subunits and the associated linker proteins (CpcA, CpcB, CpcC, CpcD, and CpcG), the reaction center polypeptides of PS I (PsaA and PsaB), and polypeptides involved in chlorophyll biosynthesis (ChlL and ChlN); many of these changes were subsequently verified by real-time RT-PCR . Additionally, the dspA mutant showed increased levels of transcripts for some genes encoding polypeptides associated with oxidative stress, such as glutathione peroxidase . Finally, additional increases in at least some of these transcripts (e.g., psbAII/III and hliA, hliB, hliC, and hliD) were observed upon exposure of the dspA mutant to HL . The transcript accumulation data presented in this paper, along with the phenotypes determined for the dspA mutant (42, 58, 73, 77), demonstrate that the mutant strain is constitutively derepressed for many genes associated with stress conditions . The DspA protein may be directly or indirectly involved in negative regulation of stress-responsive genes or the dspA mutant may be experiencing stress even under normal growth conditions . The responses of at least some of these genes that appear to be derepressed in the dspA mutant still show some response to HL (although the extent of the response might be reduced), suggesting that other mechanisms may also function to control the expression levels of these genes . However, the results from the studies in this paper do suggest that the DspA protein is likely to be a global regulatory element that either directly or indirectly controls numerous genes in the cell and may in fact be key for maintaining an appropriate homeostatic cellular milieu as environmental conditions fluctuate . Of course, there are transcripts whose abundances in wild-type cells and the dspA mutant cannot be easily explained by the hypothesis that the mutant is more sensitive to light levels . Therefore, it is likely that the DspA regulatory activity integrates with other regulatory systems to generate a final homeostatic output . It is important to understand how DspA might be sensing changing environmental conditions; the most obvious possibilities are that DspA monitors intracellular redox levels, the concentration of reactive oxygen species, and possibly even levels of blue or UV-A light (or a combinations of these factors) (77) . Other sensors thought to participate in redox control of gene expression in Synechocystis sp . strain PCC6803 have recently been identified by Li and Sherman (52) .
The dspA mutant appeared to be constitutively activated for many genes that are thought to be constituents of the pho regulon . These components include pstS, pstC, a gene encoding an alkaline phosphatase, and nucH . However, many of these genes are also activated under HL conditions . Interestingly, the sensory polypeptide SphS (sll0337) of Synechocystis sp . strain PCC6803 may contain a PAS domain (4), which has the potential to bind a chromophore and sense light and/or redox levels and coordinate light availability with the acquisition and assimilation of phosphate . Alternatively, it is possible that a major function of DspA is to manage the phosphate budget of the cell, and when this is not properly achieved in the mutant strain, the pho regulon is activated and many genes that are generally associated with stress and the absorption of excess excitation energy become active . Secondary effects may also cause the activation of some of the ccm genes when the mutant is exposed to HL (the mutant cannot take up enough for optimal CO2 fixation) . All of these results raise the question, what are the primary and secondary effects of deleting dspA in Synechocystis sp . strain PCC6803? It is clear that the phenotype of the mutant and its pattern of transcript accumulation strongly suggest that it is experiencing generalized stress responses that occur even under LL and high-nutrient conditions; these responses in many cases match the responses observed for wild-type cells transferred to HL . However, it is still not clear which genes are under the direct control of DspA as a consequence of its interactions with specific response regulators and exactly what factor(s) in the intracellular environment is being monitored by this sensory polypeptide .
| ACKNOWLEDGMENTS |
|---|
We thank Bi-Huei Hou, Katrina Ramonell, and Sue Thayer in Shanna
Somerville's laboratory for providing technical support and useful
discussion on microarray spotting and processing . We are grateful for
the help provided by Janos Demeter in the Department of Genetics and
the Stanford Microarray Database (SMD) for microarray data
normalization, retrieval, and analysis . We specially thank Q . He for
his helpful suggestions during the course of this work .
This work was supported by funds awarded to A.R.G . by the Carnegie Institution of Washington, NSF (grant MCB9727836), and USDA (grant USDA98-35301) .
| FOOTNOTES |
|---|
* Corresponding author . Mailing address: Department of Plant
Biology, The Carnegie Institution of Washington, 260 Panama St., Stanford, CA
94305 . Phone: (650) 325-1521, ext . 331 . Fax: (650) 325-6857 . E-mail: cjtu@stanford.edu .
This is manuscript no . 1632 from the Carnegie Institution .
| REFERENCES |
|---|
-subunit
phycocyanobilin lyase . J . Biol . Chem . 269:8686-8694 .
-subunit
phycocyanobilin lyase . Proc . Natl . Acad . Sci . USA 89:7017-7021.
