UV hyper-resistance in Prochlorococcus MED4 results
from a single base pair deletion just upstream of an
operon encoding nudix hydrolase and photolyase
The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.
Citation
Osburne, Marcia S. et al. “UV hyper-resistance in
Prochlorococcus MED4 results from a single base pair deletion
just upstream of an operon encoding nudix hydrolase and
photolyase.” Environmental Microbiology 12.7 (2010): 19781988. © 2010 Society for Applied Microbiology and Blackwell
Publishing Ltd
As Published
http://dx.doi.org/10.1111/j.1462-2920.2010.02203.x
Publisher
Society for Applied Microbiology / Blackwell
Version
Original manuscript
Accessed
Thu May 26 18:24:02 EDT 2016
Citable Link
http://hdl.handle.net/1721.1/61006
Terms of Use
Attribution-Noncommercial-Share Alike 3.0
Detailed Terms
http://creativecommons.org/licenses/by-nc-sa/3.0/
rP
Fo
UV hyper-resistance in Prochlorococcus MED4 results from a
single base pair deletion just upstream of an operon
encoding nudix hydrolase and photolyase
Journal:
Environmental Microbiology and Environmental Microbiology
Reports
ee
Manuscript ID:
Manuscript Type:
Date Submitted by the
Author:
Environmental Microbiology
ev
Complete List of Authors:
EMI - Research article
rR
Journal:
Draft
w
ie
Osburne, Marcia; Massachusetts Institute of Technology, Civil and
Environmental Engineering
Holmbeck, Brianne; Massachusetts Institute of Technology, Civil
and Environmental Engineering
Frias-Lopez, Jorge; The Forsyth Institute
Steen, Robert; Harvard University, Biopolymers Facility
Huang, Katherine; Massachusetts Institute of Technology, Civil and
Environmnetal Engineering
Kelly, Libusha; Massachusetts Institute of Technology, Civil and
Environmental Engineering
Coe, Allison; Massachusetts Institute of Technology, Civil and
Environmental Engineering
Waraska, Kristin; Harvard University, Biopolymers Facility
Gagne, Andrew; Harvard University, Biopolymers Facility
Chisholm, Sallie; Massachusetts Institute of Technology, Dept(s) of
Civil Environmental Engineering and Biology
ly
On
Keywords:
bacteria, evolution/evolutionary processes/gene transfer/mutation,
gene expression/regulation, genome replication/repair, growth and
survival, microbial ecology, microbial genetics, viruses
Wiley-Blackwell and Society for Applied Microbiology
Page 43 of 85
1
UV hyper-resistance in Prochlorococcus MED4 results from a single base
2
pair deletion just upstream of an operon encoding nudix hydrolase and
3
photolyase
4
Marcia S. Osburne1*, Brianne M. Holmbeck1, Jorge Frias-Lopez1†, Robert
6
Steen2, Katherine Huang 1, Libusha Kelly1, Allison Coe1, Kristin Waraska2,
7
Andrew Gagne2, and Sallie W. Chisholm1
8
9
1
rP
Fo
5
Department of Civil and Environmental Engineering, Massachusetts Institute of
ee
10
Technology, Cambridge, MA 02139, USA.
11
2
12
†
Biopolymers Facility, Harvard Medical School, Boston, MA 02115, USA.
rR
Current address: The Forsyth Institute, 140 The Fenway, Boston MA, 02115
13
ev
*Corresponding author:
15
Marcia S. Osburne
16
MIT, Department of Civil and Environmental Engineering
17
Room 48-321
18
15 Vassar St.
19
Cambridge, MA 02139
20
(617)-253-3310 (tel)
21
(617)258-7009 (FAX)
22
mosburne@mit.edu
w
ie
14
ly
On
23
24
Running title: UV hyper-resistant mutants of Prochlorococcus
25
Wiley-Blackwell and Society for Applied Microbiology
1
Page 44 of 85
1
2
I. Summary
Exposure to solar radiation can cause mortality in natural communities of
pico-phytoplankton, both at the surface and to a depth of at least 30 m. DNA
4
damage is a significant cause of death, mainly due to cyclobutane pyrimidine
5
dimer formation, which can be lethal if not repaired. While developing a UV
6
mutagenesis protocol for the marine cyanobacterium Prochlorococcus, we
7
isolated a UV hyper-resistant variant of high light-adapted strain MED4. The
8
hyper-resistant strain was constitutively upregulated for expression of the mutT-
9
phrB operon, encoding nudix hydrolase and photolyase, both of which are
rP
Fo
3
involved in repair of DNA damage that can be caused by UV light. Photolyase
11
(PhrB) breaks pyrimidine dimers typically caused by UV exposure, using energy
12
from visible light in the process known as photoreactivation. Nudix hydrolase
13
(MutT) hydrolyzes 8-oxo-dGTP, an aberrant form of GTP that results from
14
oxidizing conditions, including UV radiation, thus impeding mispairing and
15
mutagenesis by preventing incorporation of the aberrant form into DNA. These
16
processes are error-free, in contrast to error-prone SOS dark repair systems that
17
are widespread in bacteria. The UV hyper-resistant strain contained only a single
18
mutation: a 1-base pair deletion in the intergenic region directly upstream of the
19
mutT-phrB operon. Two subsequent enrichments for MED4 UV hyper-resistant
20
strains from MED4 WT cultures gave rise to strains containing this same 1-base
21
pair deletion, affirming its connection to the hyper-resistant phenotype. These
22
results have implications for Prochlorococcus DNA repair mechanisms, genome
23
stability, and possibly lysogeny.
w
ie
ev
rR
ee
10
ly
On
24
25
Wiley-Blackwell and Society for Applied Microbiology
2
Page 45 of 85
1
2
II. Introduction
Prochlorococcus, a marine cyanobacterium that numerically dominates the
mid-latitude oceans, is the smallest known oxygenic phototroph and plays a
4
central role in the oceanic carbon cycle. Its numbers can reach as high as 105
5
cells ml-1. The group consists of genetically distinct ecotypes that can be broadly
6
classified as high or low light-adapted, depending on their optimal light intensity
7
for growth (reviewed in Coleman and Chisholm, 2007). The high light-adapted
8
ecotypes are found throughout the euphotic zone of the oceans (down to a depth
9
of 150-200m), reaching highest abundance in surface waters, whereas their low
11
light-adapted counterparts dominate deeper waters.
ee
10
rP
Fo
3
Factors that affect mortality of Prochlorococcus include grazing, mainly
rR
from phagotrophic flagellates and microzooplankton (Guillou et al., 2001; Christaki
13
et al., 2002; Frias-Lopez et al., 2008), and phage infection (Sullivan et al., 2003)
14
In addition, for cells close to the surface, exposure to high solar radiation,
15
especially in the UVB range (280-320nm), also plays a role in mortality both at the
16
surface and to a depth of 10 - 30 m (Llabrés and Agustí, 2006; Agustí and
17
Llabrés, 2007). DNA damage, mainly due to the formation of cyclobutane
18
pyrimidine dimers, can be lethal if not repaired and is a primary cause of UV-
19
induced mortality (Sancar, 2000). The high light-adapted Prochlorococcus
20
ecotypes that dominate surface waters have very small (1.66-1.74 Mb),
21
streamlined, non-redundant genomes (Strehl et al., 1999; Kettler et al., 2007),
22
making them potentially less able to withstand the effects of multiple mutations
23
resulting from high doses of UV light. It is of interest, therefore, to determine
24
whether the high and low light-adapted ecotypes have different susceptibilities to
25
UV damage, and the mechanisms that might mediate these susceptibilities.
w
ie
ev
12
ly
On
Wiley-Blackwell and Society for Applied Microbiology
3
Page 46 of 85
1
In the process of developing a mutagenesis protocol for Prochlorococcus
using UV light (Osburne, Holmbeck, et al., in preparation) we isolated a UV hyper-
3
resistant strain (MED4 UVR1) of high light-adapted strain Prochlorococcus MED4.
4
We examined the differences between the hyper-resistant and wild type (WT)
5
cells with respect to levels of UV resistance, gene expression, and genome
6
sequence. The results allowed us to propose a mechanism for UV hyper-
7
resistance in MED4 UVR1, and to hypothesize mechanisms by which MED4
8
copes with UV light in its natural environment.
9
w
ie
ev
rR
ee
rP
Fo
2
ly
On
Wiley-Blackwell and Society for Applied Microbiology
4
Page 47 of 85
1
III. Results.
2
Isolation and initial phenotypic characterization of MED4 UV-hyper-resistant
3
strains
4
A culture of high light-adapted Prochlorococcus strain MED4 was
subjected to 10 rounds of UV treatment in which cells were grown to mid-log
6
phase, treated with UV light (Experimental Procedures), and allowed to recover
7
under standard growth conditions in visible light. Recovery of culture fluorescence
8
after UV treatment occurred more readily with increasing rounds of treatment.
9
After 7-8 rounds of treatment and transfer, the cultures were significantly more
rP
Fo
5
resistant to UV light than the parent strain, as reflected in the selection for the UV
11
hyper-resistant strain MED4 UVR1,(Fig. 1). This general pattern was also seen
12
during similar enrichment selections (not shown) for 2 additional UV hyper-
13
resistant strains (MED4 UVR2 and MED4 UVR3), carried out at a later date. To
14
confirm that the decrease in fluorescence caused by UV treatment was due to cell
15
lysis, we showed that culture fluorescence of WT and hyper-resistant cells
16
remained proportional to cell number for both UV treated and untreated cells
17
(Suppl. Fig.1),
w
ie
ev
rR
ee
10
UV resistance
19
Cultures of MED4 WT and MED4 UVR1 were grown and tested for relative
On
18
sensitivity to UV light at three different wavelengths: 254, 302, and 365 nm (UVC,
21
UVB, and UVA, respectively). At equivalent doses of UV light, MED4 UVR1 (Fig.
22
2, D, E, F) was considerably more resistant to both UVC and UVB than MED4 WT
23
(Fig. 2, A, B, C). However, both the hyper-resistant and WT strains were equally
24
insensitive to UVA at the doses used in this experiment. While UVC and UVB are
25
known to be absorbed by and cause direct damage to DNA, UVA is poorly
ly
20
Wiley-Blackwell and Society for Applied Microbiology
5
Page 48 of 85
1
absorbed by DNA, and its effect on cells is thought to be more complex and
2
indirect (Kidambi et al., 1996)
3
Growth rates of MED4 WT and MED4 UVR1
To assess whether acquisition of the UV hyper-resistant phenotype
5
affected the growth rate under standard conditions in the absence of UV light, we
6
measured the growth rates of WT and resistant strains in exponential phase. For
7
MED4 WT, µ was 0.64 +/- 0.01 day-1, compared with 0.67 +/- 0.01 day-1 for MED4
8
UVR1, a significant difference according to the two tailed t-test using independent
9
samples (P<0.05). When the experiment was repeated the growth rates were not
10
significantly different, leading us to conclude that any such difference between the
11
two strains under these conditions it is small, and can only be revealed by very
12
precise and repeated growth rate measurements. We cannot rule out that the
13
growth rates of these strains may well vary significantly under conditions prevalent
14
in their natural environment, where UV light is frequently present and other
15
selection pressures may apply.
ie
ev
rR
Gene expression
w
17
ee
16
rP
Fo
4
To begin to explore the underlying mechanisms mediating UV hyperresistance in MED4 UVR1 we compared the complete transcriptomes of MED4
19
WT and MED4 UVR1 cells in mid log-phase under standard growth conditions in
20
the absence of UV treatment, using whole genome microarrays. The only
21
significant difference seen was in the expression of two genes, phrB and mutT,
22
encoding photolyase and nudix hydrolase, respectively, which were more highly
23
expressed in MED4 UVR1 relative to MED4 WT (Fig. 3). Interestingly, these two
24
genes form a small predicted operon in MED4 WT (Fig. 6), consistent with their
25
apparent co-regulation in MED4 UVR1. These genes, including their order and
26
context between the folK and degT genes, are conserved in all the sequenced
ly
On
18
Wiley-Blackwell and Society for Applied Microbiology
6
Page 49 of 85
1
high light-adapted strains of Prochlorococcus, and also in strains NATL1A and
2
NATL2A, which show increased fitness in variable light.
3
(http://www.microbesonline.org, Kettler et al., 2007).
4
We next used qRT-PCR to analyze a time course of expression of several
selected genes following UV treatment, guided in our choice of genes by the
6
expression results from untreated MED4 WT and MED4 UVR1 cells described
7
above. In addition to mutT and phrB, already shown to be constitutively
8
upregulated in MED4 UVR1 cells, we examined the expression of DNA repair-
9
related genes recA and radA, that might also be involved in the cellular response
rP
Fo
5
to UV light: recA is known to play a pivotal role in response to bacterial DNA
11
damage (Michel, 2005), and while the recA analog radA is of uncertain function in
12
bacteria, it is thought to be involved in recombinational repair of DNA at replication
13
forks blocked due to DNA damage (possibly resulting from pyrimidine dimer
14
formation (Beam et al., 2002)). radA is also known to be upregulated in response
15
to UVB and UVC radiation in the archaeon Halobacterium sp. NRC-1 (Boubriak et
16
al., 2008).
w
ie
ev
rR
ee
10
In the absence of UV treatment, in steady state growth, MED4 UVR1
18
again had considerably higher levels of phrB and mutT transcripts relative to
19
MED4 WT (Fig. 4A), in addition to slight constitutive upregulation of radA that was
20
below the level of detection in the microarray experiment. Under the same
21
conditions, recA expression was slightly lower in MED4 UVR1 as compared with
22
MED4 WT, possibly a consequence of the constitutively high level of phrB and
23
mutT expression (for example, excessive amounts of PhrB and MutT may reduce
24
background levels of DNA damage, thus decreasing the need for RecA function).
ly
On
17
25
To explore the dynamics of the expression of these genes in response to
26
UV exposure, we compared transcript levels of UV treated and untreated MED4
Wiley-Blackwell and Society for Applied Microbiology
7
Page 50 of 85
WT and MED4 UVR1 cells 5 and 60 minutes after treatment (Figs. 4B and C). For
2
MED4 WT, phrB, mutT, and radA were rapidly upregulated, showing elevated
3
transcript levels by 5 minutes post-treatment. By 60 minutes post-treatment the
4
levels dropped – in two cases below those of the untreated cells, suggesting the
5
need for their function had diminished. In contrast, for MED4 UVR1, phrB, mutT,
6
and radA were only slightly upregulated following UV irradiation, with transcripts
7
decreasing only slightly between 5 and 60 min, consistent with their constitutively
8
elevated levels. The kinetics of upregulation of the phrB-mutT operon in the WT
9
strain (which, although rapid, takes some time as compared with constitutive
rP
Fo
1
expression in the mutant) provides an explanation for increased UV resistance in
11
the mutant strain, i.e., its constitutively upregulated enzymes are already present
12
and poised to repair damage before the onset of UV treatment, whereas in the
13
WT these enzymes need to be expressed and synthesized following UV
14
treatment, potentially allowing more unrepaired DNA damage to accumulate.
ev
rR
On the other hand, for both mutant and WT strains, recA gene expression
ie
15
ee
10
was only slightly upregulated by 5 min post-treatment, but by 60 min was present
17
at significantly higher levels than in untreated controls. The kinetics of recA
18
upregulation in both strains suggests that recA is likely to be normally regulated
19
independently of the phrB-mutT operon.
20
Sensitivity of low light-adapted Prochlorococcus MIT9313 to UV light
ly
On
21
w
16
With the knowledge that low light-adapted strains of Prochlorococcus
22
encode mutT but not phrB, and thus cannot synthesize photolyase (Kettler et al.,
23
2007), we tested the sensitivity of a low light-adapted strain (MIT9313) to UV
24
exposure and compared it with MED4 WT (Fig.5). As expected, MIT9313 was in
25
fact considerably more sensitive to UVB and UVC radiation than was MED4 WT.
26
This observation supports the contention that the photoreactivation system plays
Wiley-Blackwell and Society for Applied Microbiology
8
Page 51 of 85
a significant role in UV resistance in the ocean, and suggests that the presence of
2
phrB was selected for in the high light-adapted strain MED4 because it is a high-
3
light adapted strain that occupies high light surface waters where UV radiation is
4
present and can act as a selective agent. In contrast, photoreactivation repair
5
would not be necessary in low light-adapted strains such as MIT9313, which
6
occupy deeper, low light waters. The larger difference between the two strains
7
seen with UVB radiation than with UVC radiation may reflect the fact that UVB
8
radiation is generally less damaging than UVC, potentially allowing smaller
9
dosage differences to be titrated more precisely.
The lack of differential sensitivity to UVA treatment at the doses used is
ee
10
rP
Fo
1
consistent with the inability of UVA radiation to cause damage that can be
12
repaired by photolyase (Sancar, 2004)
13
Genome comparison of MED4 WT and MED4 UVR1
ev
14
rR
11
To determine the genomic underpinnings of the UVR phenotype we
sequenced the genomes of both MED4 WT and MED4 UVR1. Genomes were
16
first assembled using the sequence of a parent culture of MED4 (GenBank
17
BX548174) as the reference, and then confirmed using a modified de novo
18
assembly capable of detecting large insertions or deletions (Experimental
19
Procedures). MED4 WT contained 16 SNPs as compared with the published
20
reference sequence (Rocap et al., 2003, white rows in Table 2). The stability of
21
the MED4 WT genome over more than 5 years of growth and serial transfers in
22
the laboratory seems remarkable.
w
ly
On
23
ie
15
Relative to the reference sequence, MED4 UVR1 contained the same 16
24
SNPs as MED4 WT; in addition, however, it contained a single base pair deletion
25
(shaded row, Table 2), located just upstream of the mutT gene, in a putative
26
regulatory region of the mutT-phrB operon (Fig. 6). This result was confirmed by
Wiley-Blackwell and Society for Applied Microbiology
9
Page 52 of 85
1
Sanger sequencing of a PCR fragment encompassing the 1 bp deletion
2
(Experimental Procedures), and strongly suggests that this mutation, located in a
3
possible regulatory region for the downstream operon, is responsible for the high-
4
level constitutive expression of the operon in the mutant, thus conferring UV-
5
hyper-resistance.
Because we do not have robust genetic tools for Prochlorococcus, we
rP
Fo
6
were unable to test directly whether this deletion would give rise to the UV-hyper-
8
resistant phenotype when transferred to a naïve WT strain. To further
9
substantiate our hypothesis, therefore, we sequenced this region of the genome
10
(containing the single base pair deletion in MED4 UVR1) in the other UV hyper-
11
resistant strains, MED4 UVR2 and MED4 UVR3. Remarkably, both MED4 UVR2
12
and MED4 UVR3 carried the identical 1 bp deletion upstream of the mutT-phrB
13
operon (Fig. 6). Although the mechanism by which this sequence may control
14
expression of the operon is not clear, this result provides compelling evidence that
15
the deletion is responsible for the UVR phenotype, as the strong UV challenge
16
repeatedly enriched for cells containing the identical deletion mutation. Although
17
there are no obvious regulatory motifs present in the intergenic sequence, and
18
such sequences have not yet been defined for Prochorococcus, a conceivable
19
scenario is that a transcriptional activator or repressor might normally bind to the
20
upstream region to regulate expression of the mutT-phrB operon in the WT case.
21
The deletion in the mutant strain may disrupt the binding site, either preventing
22
binding of the repressor, or causing the activator to bind more efficiently, resulting
23
in constitutive expression. The role of radA in this model is unclear, as the
24
function of its encoded enzyme remains to be elucidated. Additional studies of
25
global transcription following UV treatment should clarify these issues further.
w
ie
ev
rR
ee
7
ly
On
Wiley-Blackwell and Society for Applied Microbiology
10
Page 53 of 85
1
We note that MED4 UVR1 has not reverted to UV-sensitivity despite being
maintained under non-selective conditions for several years, which might be
3
explained both by the fact that spontaneous compensatory 1 bp insertion
4
mutations are relatively rare (Schaaper and Dunn, 1991; Mira et al., 2001), and
5
that MED4 WT does not appear to grow faster than the mutant under laboratory
6
conditions.
7
Analysis of the mutT genomic region in sequences from wild Prochlorococcus
8
9
rP
Fo
2
Because increased resistance to UV radiation could conceivably confer an
advantage to Prochlorococcus that dominate surface waters, it was of interest to
ee
determine whether the mutant sequence could be found in the wild. We
11
searched the GOS (Global Ocean Survey) metagenomic database (Experimental
12
Procedures) to see whether we could find any incidence of the single base
13
deletion in fragments of Prochlorococcus genomes that could be identified as
14
coming from directly upstream of mutT. The search revealed 31 sequences with
15
high similarity to MED4, one of which contained the SNP found in our mutant
16
strains (JCVI_READ_1103359078035). 7 additional sequences contained “G” in
17
place of “A” at the same position as the deletion mutation (i.e., AAAAGA instead
18
of AAAA_A (mutant) or AAAAAA (WT) – see Table 3). The remaining 23
19
sequences contained the WT sequence AAAAAA at the same location (Table 3).
20
While there is not enough data to make statistical arguments, and we have not
21
taken sequencing error into consideration – especially for this homopolymeric
22
stretch of DNA – we find these data provocative albeit inconclusive with regard to
23
the relevance of our laboratory observations to wild populations of
24
Prochlorococcus. We also emphasize that the number of sequences analyzed is
25
vanishingly small relative to the 104 – 105 Prochlorococcus cells ml-1 populating
26
many of the stations sampled in GOS. Thus even if the mutation were present at
w
ie
ev
rR
10
ly
On
Wiley-Blackwell and Society for Applied Microbiology
11
Page 54 of 85
1
a high frequency, we would not necessarily expect to see it in a sample of this
2
size.
w
ie
ev
rR
ee
rP
Fo
ly
On
Wiley-Blackwell and Society for Applied Microbiology
12
Page 55 of 85
1
IV. Discussion
Using visible light, especially in the blue-violet range, as a source of
3
energy, photolyases catalyze the repair of cyclobutane dipyrimidine dimers
4
formed by UV damage to DNA (Sancar, 2004). Metagenomics studies have
5
shown that photolyase and cryptochrome gene homologs have been found to be
6
highly represented in surface seawater (DeLong et al., 2006; Frias-Lopez et al.,
7
2008; Singh et al., 2009), in one instance, representing ~0.1% of the genes
8
present in a sample from a high-UV ocean environment (Singh et al., 2009). Such
9
environments are also enriched for transcripts of these genes (Frias-Lopez et al.,
rP
Fo
2
2008). Results of these studies are consistent with the high levels of UV radiation
11
to which the ocean surface is exposed daily, and which, according to recent
12
measurements, may penetrate to depths of tens of meters (Tedetti and Sempere,
13
2006). Moreover, photolyase is encoded on the genomes of only a subset of the
14
sequenced Prochlorococcus strains, specifically those that are high light-adapted
15
or display increased relative fitness in fluctuating light (Kettler et al., 2007). Thus
16
photolyase, specifically needed to repair radiation damage to DNA, appears to be
17
maintained when there is a strong selection for its activity. This strong selection is
18
reflected by the increased sensitivity to UVB and C of low light-adapted strain
19
MIT9313, which does not encode phrB, as compared with MED4 WT. Further,
20
this difference in sensitivity between MIT9313 and MED4 WT was not observable
21
in response to UVA radiation, which does not cause the formation of pyrimidine
22
dimers. Whereas UBC and UVB can cause direct damage to DNA, UVA (320 to
23
400 nm) causes only indirect damage through reactive oxygen intermediates
24
(Kidambi et al., 1996) .
w
ie
ev
ly
On
26
rR
25
ee
10
Nudix hydrolase, on the other hand, is encoded by the “core” gene mutT,
found in all the Prochlorococcus genomes sequenced to date (Kettler et al.,
Wiley-Blackwell and Society for Applied Microbiology
13
Page 56 of 85
2007). This is consistent with its function to repair damage to GTP caused by
2
reactive oxygen species formed not only as a result of oxidizing radiation such as
3
UV light, but also formed as a result of oxidative stress from activities such as
4
photosynthesis (Latifi et al., 2009). It is noteworthy that although UVC radiation
5
does not normally penetrate the earth’s atmosphere, the levels of UVA and UVB
6
radiation measured at the surface of the ocean and possibly tens of meters below
7
may reach or greatly exceed the doses used in these experiments (Buma et al.,
8
2001).
9
rP
Fo
1
Interestingly, photolyase and nudix hydrolase repair activities are not error
prone and are therefore non-mutagenic, in contrast to SOS repair (a form of dark
11
repair) that allows survival at the cost of mutagenesis. Despite the fact that MED4
12
and other strains of Prochlorococcus encode a number of SOS-type genes
13
including recA and lexA (the canonical repressor of SOS genes), the potential
14
advantage of using error-free repair is apparent, as it would seem that the small,
15
streamlined genomes of the high light-adapted ecotypes of Prochlorococcus
16
would be adversely affected by high mutation rates. In support of this idea,
17
Prochlorococcus strains also lack low-fidelity DNA polymerases II and IV, which
18
are known to replicate across lesions, thereby leading to mutations (Napolitano et
19
al., 2000; Michel, 2005). They do, however, encode one of the low-fidelity DNA
20
polymerases, PolV (encoded by umuCD). It is noteworthy that selection for
21
resistance to repeated UV light exposure in E. coli has produced mutants that are
22
typically affected in several different “dark repair” genes involved in DNA
23
replication and repair (Alcantara-Diaz et al., 2004), including radA uvrA, and polA,
24
whereas enrichment for UV hyper-resistance in MED4 resulted in a lesion
25
affecting photoreactivation repair. Although the specific E. coli UV resistance
26
mutations isolated may at least partly be an artifact of their selection, i.e., they
w
ie
ev
rR
ee
10
ly
On
Wiley-Blackwell and Society for Applied Microbiology
14
Page 57 of 85
1
were selected in the dark in order to avoid photoreactivation repair following
2
mutagenesis, it is tempting to speculate that with a larger and less streamlined
3
genome, E. coli might be better able to tolerate mutations caused by error-prone
4
repair mechanisms and may therefore make more use of these mechanisms.
5
Photolyase has been shown to be the major UV-resistance factor for
another cyanobacterium, Synechocystis PCC 6803 (Ng and Pakrasi, 2001). This
7
was shown genetically, i.e., a knockout of the phrB gene conferred increased UV
8
sensitivity to the organism, whereas a recA knockout mutant was still relatively
9
resistant to UVC (Minda et al., 2005). The authors suggested that RecA may be
10
used primarily for other functions in this organism, such as recombination repair,
11
which is error-free (Sargentini and Smith, 1985; Cox, 1999). Although
12
Synechocystis PCC 6803 also encodes a lexA gene, it may also be used for
13
functions other than the regulation of SOS repair. Our data show that the
14
Prochlorococcus recA gene was upregulated in both MED4 WT and MED4 UVR1
15
following UV irradiation, but only after the phrB and mutT genes were
16
upregulated. It is not known whether MED4 or any Prochlorococcus strain makes
17
use of an SOS repair system, and if so, under what circumstances. However,
18
regardless of whether the 1-bp deletion was a new mutation caused by UV
19
treatment or a pre-existing mutant in the original MED4 batch culture, it is striking
20
that three enrichments for UV hyper-resistance , carried out at different times from
21
the same MED4 WT strain using 10 rounds of treatment with UVC radiation of
22
sufficient intensity to kill most of the cells, selected for this same 1 bp deletion
23
mutation - perhaps suggesting there may be limited ways of becoming UV-hyper-
24
resistant. Further, a comparison of the genome sequences of MED4 WT and
25
MED4 UVR1 points to high genome stability in the surviving cells. Again, with
26
frequent and long-lasting environmental UV irradiation exposure, it would seem
w
ie
ev
rR
ee
rP
Fo
6
ly
On
Wiley-Blackwell and Society for Applied Microbiology
15
Page 58 of 85
1
advantageous for Prochlorococcus to make use primarily of error-free
2
photoreactivation repair systems, especially with the abundance of visible light
3
available to power the system.
4
The role of an SOS repair system in Prochlorococcus has yet to be
established. Future experiments that measure expression of relevant candidate
6
SOS and other DNA repair genes under conditions designed to simulate
7
exposures in the natural environment will address this question. In this context it
8
is interesting to note that true temperate phage that can be excised from the
9
bacterial genome require SOS-type processes, initiated by DNA damage, for their
rP
Fo
5
excision (Rokney et al., 2008). To date, despite the isolation of many
11
Prochlorococcus cyanophage and despite strong evidence of the presence of
12
phage genes in both low and high light-adapted Prochlorococcus genomes,
13
temperate phage have not yet been isolated for this organism.. An increased
14
understanding of the role of SOS and photoreactivation in Prochlorococcus DNA
15
damage repair in low and high light-adapted strains may help to elucidate
16
possibilities for lysogeny in Prochlorococcus.
w
ie
ev
rR
ee
10
17
ly
On
Wiley-Blackwell and Society for Applied Microbiology
16
Page 59 of 85
1
V. Experimental Procedures.
2
Growth conditions
3
All Prochlorococcus cultures were grown at 22°C under constant light
provided by cool-white fluorescent lamps at irradiances of ~ 25 mol Q m−2 s−1 for
5
MED4 strains, and ~ 16 mol Q m−2 s for MIT9313. Cultures were grown in Pro99
6
medium, consisting of sterile (0.2 µm filtered, autoclaved) Sargasso Sea water
7
supplemented with 800 µM NH4Cl, 50 µM NaH2PO4, and trace metals (Moore et
8
al., 2007). Growth was monitored using Turner Design fluorometer 10-AU to
9
measure bulk chlorophyll fluorescence, used as a proxy for culture cell density
rP
Fo
4
(Moore et al., 2007). Bulk chlorophyll fluorescence has been shown to be directly
11
proportional to cell concentration for cells growing in exponential phase (Moore et
12
al., 2007), as determined by measuring cell number throughout log phase by
13
means of flow cytometry.
14
UV treatment and enrichment for hyper-resistant mutants
ev
rR
In all cases, UV treatment was applied using a UVLMS-38 8-W UV lamp
ie
15
ee
10
(UVP, Upland, CA), generating wavelengths of 365/302/254 nm, with cells at a
17
distance of 22 cm below the light source. Cells were slowly rotated using a
18
mechanical rotator platform during all UV treatments. For resistance enrichment,
19
cells were irradiated at room temperature for 30 seconds at 254 nm (UV fluence
20
9.2 J m-2) in an uncovered sterile glass Petri dish (150 mm diameter) containing
21
25 ml of culture grown to a fluorescence of ~100 (early- to mid-log phase).
22
Following treatment, cells were immediately returned to standard growth
23
conditions. In cases where fluorescence decreased following UV treatment, cells
24
were maintained under standard growth conditions until they recovered (i.e.,
25
reached a fluorescence of 100). After recovery cells were diluted 1:10 into Pro99
26
medium and again grown to fluorescence of ~100 for the next round of UV
w
16
ly
On
Wiley-Blackwell and Society for Applied Microbiology
17
Page 60 of 85
1
irradiation. To confirm that culture fluorescence was proportional to cell number
2
following UV treatment, cell counts on selected samples were carried out by
3
means of flow cytometry, using 2 µm fluorescent beads as a size reference
4
(Moore et al., 2007). Results are shown in Suppl. Fig. 1.
Once isolated, UV-hyper-resistant strains were purified at least 20 times
6
by dilution to very low titer (10 cells/ml), followed by regrowth. This process was
7
necessary because it is difficult, and in this case not possible, to isolate and re-
8
grow pure single colonies of Prochlorococcus. In addition, at least 10 cells/ml
9
were required for these cultures to grow in liquid. However, DNA sequencing of
10
mutant vs. WT batch cultures confirmed the appropriate genotypes, i.e., all three
11
mutant cultures contained the same deletion mutation upstream of the mutT-phrB
12
operon. . Furthermore, in spite of continuous culturing, involving cycles of dilution
13
and re-growth, for periods longer than 2 years, the mutants have retained their
14
UV-hyper-resistant genotype and phenotype..
15
RNA preparation
ie
ev
rR
ee
rP
Fo
5
Cells (UV treated or untreated) were collected by centrifugation (7 min,
17
12,000 x g, 22°C), resuspended in 1 ml of RNA later (Ambion, Inc. Austin, TX,
18
USA), quick frozen and stored at -80ºC. RNA was isolated using the Mirvana
19
miRNA isolation kit (Ambion) according to the manufacturer’s instructions. DNA
20
was removed using Turbo DNase treatment (Ambion) according to the
21
manufacturer’s instructions and DNA removal was confirmed by PCR using the
22
primers shown in Table 1 for qRT-PCR. RNA was then ethanol precipitated and
23
resuspended to a concentration of 100 ng ml-1.
24
qRT-PCR reactions
ly
26
On
25
w
16
RNA was isolated as described above, reverse-transcribed, and subjected
to triplicate real-time PCR reactions as described previously (Frias-Lopez et al.,
Wiley-Blackwell and Society for Applied Microbiology
18
Page 61 of 85
2008) , using the primer sets shown in Table 1. We calculated gene expression
2
levels using the Comparative CT (2-∆∆CT) method (Livak and Schmittgem, 2001;
3
Frias-Lopez et al., 2008), where CT is the threshold cycle at which fluorescence
4
rises above background levels, and ∆∆ CT = ∆CT , sample - ∆ CT, reference.
5
Expression of gene rnpB was used as the endogenous reference.
6
Microarray hybridization and analysis
7
rP
Fo
1
For microarray hybridizataion, 100 ng of RNA was first amplified using the
8
MessageAmp II-Bacteria Kit (Ambion) according to the manufacturer’s
9
instructions, as previously described (Frias-Lopez et al., 2008).
ee
cDNA was synthesized, labeled, and hybridized to customized MD4-9313
11
microarrays (Affymetrix, Santa Clara, CA), and scanned, all as previously
12
described (Lindell et al., 2005). Hybridizations were visualized using GeneSpring
13
software (version 7.3.1; Silicon Genetics, Palo Alto, CA), initially normalized using
14
the Robust Multichip Average algorithm (Bolstad et al., 2003) implemented in
15
GeneSpring, and were later normalized using the Loess correction performed by
16
using the open-source Statistical Data Analysis software R version 2.5.0
17
(http://www.r-project.org/). We used the program QVALUE, which measures
18
significance in terms of the false discovery rate (Storey and Tibshirani, 2003). A
19
gene was considered as differentially expressed if the q value was <0.05.
20
Genome Sequencing
w
ie
ev
ly
On
21
rR
10
Library Preparation: Genomic DNA libraries suitable for sequencing on the
22
Illumina Genome Analyzer were prepared as described (manufacturer’s document
23
“11251892_Genomic_DNA_Sample_Prep.pdf,”) except that genomic DNA (30 µg
24
in 300 µl TE, in an ice bath solution) was fragmented using a model UCD-200
25
Bioruptor sonicator (Diagenode): low power, 15 seconds on/15 seconds off cycle,
Wiley-Blackwell and Society for Applied Microbiology
19
Page 62 of 85
for a total of 10 minutes. Adaptor-DNA constructs were purified on a 2% agarose-
2
TAE gel, run at 120V for 60 min, with a low molecular weight DNA ladder (New
3
England BioLabs) run in parallel. Blank wells between each sample prevented
4
DNA cross contamination. DNA of size150 -200 bp was excised from the gel and
5
purified using a QIAGEN Gel Extraction Kit. Samples were eluted into 50 µl
6
elution buffer, then extracted using the Qiagen MinElute Gel Extraction Kit and
7
eluted into 30 µl of elution buffer.
8
9
rP
Fo
1
Purifed adapter-DNA constructs were enriched for fragments having the
proper adaptor molecules on both ends using PCR (1 µl DNA, 25 µl Phusion DNA
polymerase (Finnzymes Oy), 1 µl each of PCR primers 1.1 and 2.1 (© 2006 and
11
2008 Illumina, Inc.), and 22 µl water). The PCR protocol was:30 sec at 98°C, 18
12
cycles of: 10 sec at 98°C, 30 sec at 65°C, 30 sec at 72°C, followed by a final
13
extension of 5 min at 72°C. Amplified DNA was purified using the Qiagen
14
QIAquick PCR Purification Kit, and eluted into 50 µl of elution buffer. The DNA
15
library concentration was determined using a NanoDrop ND8000
16
spectrophotometer, followed by analysis on an Agilent 2100 Bioanalyzer using the
17
Agilent RNA 6000 Pico Chip.
w
ie
ev
rR
ee
10
Sequencing Methods: DNA library samples were diluted to a
19
concentration of 10 nM, and were clustered at a final concentration of 2.0 pM
20
each on an Illumina GA flow cell, according to the manufacturer’s standard
21
protocols. Sequencing of the samples on the clustered flow cell were performed
22
on the Genome Analyzer using a single read 36 cycle protocol, following the
23
manufacturer’s standard procedures and protocols.
ly
24
On
18
Data Analysis: Each sample (MED4 WT and MED4 UVR1) was run on
25
one flow cell lane of an Illumina GA I, yielding 5.8 and 7.8 million 36-base
26
sequence reads, respectively. Raw images were processed using the Genome
Wiley-Blackwell and Society for Applied Microbiology
20
Page 63 of 85
Analyzer Pipeline Software v0.3 to produce DNA sequence files in FASTQ format.
2
Output files were imported and assembled using the SeqMan NGEN v1.2
3
program from DNAStar with the original published MED4 genome (BX548174) as
4
the reference sequence. To check for possible large insertions, deletions,
5
inversions, or genomic rearrangements, further assembly analysis was carried out
6
using NGen’s Split Reference Sequence Strategy
7
(http://www.dnastar.com/products/ReferenceGuidedAssembly.php). Resulting
8
genome assemblies were imported into DNAStar LaserGene8, where mutations
9
were identified.
11
Sequencing of PCR fragments
ee
10
rP
Fo
1
DNA fragments spanning the putative regulatory region upstream of mutT
rR
were PCR-amplified using forward and reverse primers:
13
(5’GGTGGGAAGATTTTTATACGGTTGACGAG-3’, and
14
5’-AATGCTGCGCCAGGATGC-3’), respectively. Amplified DNA fragments were
15
treated with ExoSAP-IT (USB, Cleveland, OH) and submitted for pyrosequencing
16
(Amplicon Express, Pullman, WA).
17
Analysis of GOS Database
w
ie
The GOS database was searched to recruit reads and their paired ends
On
18
ev
12
homologous to a 30 bp stretch of the 51-bp intergenic region (Fig. 6) between folK
20
and mutT. This search yielded 175 sequences that matched at least 30 bp of the
21
intergenic region. To select those with high similarity (at least 95%) to high-light
22
Prochlorococcus strains over the entire length of the read (normally ~700 – 1000
23
bp), BLASTn was used to compare the reads against the microbial genomes
24
database in Microbes Online (http://www.microbesonline.org). 31 reads had high
25
similarity to Prochlorococcus, and these were analyzed for the presence of SNPs
26
in the intergenic region, as described above. Reads of interest can be accessed
ly
19
Wiley-Blackwell and Society for Applied Microbiology
21
Page 64 of 85
1
in the CAMERA (Community Cyberinfrastructure for Advanced Marine Microbial
2
Ecology Research and Analysis) database (http://camera.calit2.net/index.php).
w
ie
ev
rR
ee
rP
Fo
ly
On
Wiley-Blackwell and Society for Applied Microbiology
22
Page 65 of 85
VI. Acknowledgements
2
We thank D. Sher, and A. Thompson for assistance with data analysis, Q. Zeng
3
and D. Sher for reviewing the manuscript, and all members of the Chisholm
4
laboratory for interesting and helpful discussions. Thanks also to J. Engelking of
5
DNAStar for his work on the genome assemblies. This research was supported by
6
grants to SWC from the Gordon and Betty Moore Foundation, the Department of
7
Energy Genomics GTL program, the National Science Foundation, and grants to
8
BMH from the Howard Hughes Medical Institute and the Lord Foundation.
w
ie
ev
rR
ee
rP
Fo
1
ly
On
Wiley-Blackwell and Society for Applied Microbiology
23
Page 66 of 85
1
VII. References
2
Agustí, S., and Llabrés, M. (2007) Solar Radiation-induced Mortality of Marine
3
Pico-phytoplankton in the Oligotrophic Ocean. Photochemistry and
4
Photobiology 83: 793-801.
5
Alcantara-Diaz, D., Brena-Valle, M., and Serment-Guerrero, J. (2004) Divergent
adaptation of Escherichia coli to cyclic ultraviolet light exposures.
7
Mutagenesis 19: 349-354.
8
9
Beam, C., Saveson, C., and Lovett, S. (2002) Role for radA/sms in recombination
intermediate processing in Escherichia coli. J Bacteriol. 184: 6836-6844.
Boelen, P., Post, A., Veldhuis, M., and Buma, A. (2002) Diel patterns of UVBR-
ee
10
rP
Fo
6
induced DNA damage in picoplankton size fractions from the Gulf of
12
Aqaba, Red Sea. Microb Ecol. 44: 164-174.
13
rR
11
Bolstad, B., Irizarry, R., Astrand, M., and Speed, T. (2003) A comparison of
ev
normalization methods for high density oligonucleotide array data based
15
on variance and bias. Bioinformatics 19: 185-193.
16
ie
14
Boubriak, I., Ng, W., Dassarma, P., S Dassar, a., Crowley, D., and McCready, S.
w
(2008) Transcriptional responses to biologically relevant doses of UV-B
18
radiation in the model archaeon, Halobacterium sp. NRC-1. Saline
19
Systems 4.
Buma, A., Helbling, E., Boer, M.d., and Villafañe, V. (2001) Patterns of DNA
ly
20
On
17
21
damage and photoinhibition in temperate South-Atlantic
22
picophytoplankton exposed to solar ultraviolet radiation. J Photochem
23
Photobiol B. 62: 9-18.
24
Christaki, U., Courties, C., Karayanni, H., Giannakourou, A., Maravelias, C.,
25
Kormas, K., and Lebaron, P. (2002) Dynamic characteristics of
26
Prochlorococcus and Synechococcus consumption by bacterivorous
Wiley-Blackwell and Society for Applied Microbiology
24
Page 67 of 85
1
2
nanoflagellates. Microb Ecol. 43: 341-352.
Coleman, M., and Chisholm, S. (2007) Code and context: Prochlorococcus
3
as a model for cross-scale biology. Trends in Microbiology 15: 398-
4
407.
5
7
rP
Fo
6
Cox, M. (1999) Recombinational DNA repair in bacteria and the RecA protein.
Prog Nucleic Acid Res Mol Biol. 63: 311-366.
DeLong, E., Preston, C., Mincer, T., Rich, V., Hallam, S., Frigaard, N. et al. (2006)
8
Community genomics among stratified microbial assemblages in the
9
ocean's interior. Science 311: 496-503.
ee
10
Frias-Lopez, J., Shi, Y., Tyson, G., Coleman, M., Schuster, S., Chisholm, S., and
Delong, E. (2008) Microbial community gene expression in ocean
12
surface waters. Proc Natl Acad Sci U S A. 105: 3805-3810.
13
rR
11
Guillou, L., Jacquet, S., Chretiennot-Dinet, M.-J., and Vaulot, D. (2001) Grazing
ev
impact of two small heterotrophic flagellates on Prochlorococcus and
15
Synechococcus. . Aquat. Microb. Ecol. 26: 201-207.
Karlin, S., and Brocchiere, L. (1996) Evolutionary conservation of RecA genes in
w
16
ie
14
relation to protein structure and function. J Bacteriol. 178: 1881-1894.
18
Kettler, G., Martiny, A., Huang, K., Zucker, J., Coleman, M., Rodrigue, S. et al.
On
17
(2007) Patterns and implications of gene gain and loss in the evolution of
20
Prochlorococcus. PLoS Genet. 3: e231.
ly
19
21
Kidambi, S., Booth, M., Kokjohn, T., and Miller, R. (1996) recA-dependence
22
of the response of Pseudomonas aeruginosa to UVA and UVB
23
irradiation. Microbiology 142: 1033-1040.
24
25
Latifi, A., Ruiz, M., and Zhang, C. (2009) Oxidative stress in cyanobacteria. FEMS
Microbiol Rev. 33: 258-278.
Wiley-Blackwell and Society for Applied Microbiology
25
Page 68 of 85
1
Lindell, D., Jaffe, J., Johnson, Z., Church, G., and Chisholm, S. (2005)
2
Photosynthesis genes in marine viruses yield proteins during host
3
infection. Nature 438: 86-89.
4
Livak, K., and Schmittgem, T. (2001) Analysis of relative gene expression
data using real-time quantitative PCR and the 2−∆∆CT method. .
6
Methods 25: 402-408.
7
rP
Fo
5
Llabrés, M., and Agustí, S. (2006) Picophytoplankton cell death induced by UV
8
radiation: Evidence for oceanic Atlantic communities. Limnology and
9
Oceanography. 51: 21-29.
ee
Matallana-Surget, S., Meador, J., Joux, F., and Douki, T. (2008) Effect of the GC
11
content of DNA on the distribution of UVB-induced bipyrimidine
12
photoproducts. Photochemical and Photobiological Sciences 7: 794-801.
14
17
Michel, B. (2005) After 30 years of study, the bacterial SOS response still
surprises us. PLoS Biol. 3: e255.
w
16
143.
ie
15
McLennan, A. (2006) The Nudix hydrolase superfamily. Cell Mol Life Sci. 63: 123-
ev
13
rR
10
Minda, R., Ramchandani, J., Joshi, V., and Bhattacharjee, S. (2005) A
On
homozygous recA mutant of Synechocystis PCC6803: construction
19
strategy and characteristics eliciting a novel RecA independent UVC
20
resistance in dark. Mol Genet Genomics 274: 616-624.
21
22
23
ly
18
Mira, A., Ochman, H., and Moran, N. (2001) Deletional bias and the evolution of
bacterial genomes. Trends Genet. 17: 589-596.
Moore, L., Coe, A., Zinser, E., Saito, M., Sullivan, M., Lindell, D. et al. (2007)
24
Culturing the marine cyanobacterium Prochlorococcus. Limnol.
25
Oceanogr.: Methods 5: 353-362.
Wiley-Blackwell and Society for Applied Microbiology
26
Page 69 of 85
1
Napolitano, R., Janel-Bintz, R., Wagner, J., and Fuchs, R. (2000) All three SOS-
2
inducible DNA polymerases (Pol II, Pol IV and Pol V) are involved in
3
induced mutagenesis. EMBO J. 19: 6259-6265.
4
Ng, W., and Pakrasi, H. (2001) DNA photolyase homologs are the major UV
resistance factors in the cyanobacterium Synechocystis sp. PCC 6803.
6
Mol Gen Genet. 264: 924-930.
7
rP
Fo
5
Rocap, G., Larimer, F., Lamerdin, J., Malfatti, S., Chain, P., Ahlgren, N. et
8
al. (2003) Genome divergence in two Prochlorococcus ecotypes
9
reflects oceanic niche differentiation. Nature 424: 1042-1047.
ee
10
Rokney, A., Kobiler, O., Amir, A., Court, D., Stavans, J., Adhya, S., and
Oppenheim, A. (2008) Host responses influence on the induction of
12
lambda prophage. Mol Microbiol. 68: 29-36.
13
17
19
21
Sargentini, N., and Smith, K. (1985) Spontaneous mutagenesis: the roles of DNA
repair, replication, and recombination. Mutat Res 154: 1-27.
Schaaper, R., and Dunn, R. (1991) Spontaneous mutation in the Escherichia coli
lacI gene. Genetics 129: 317- 326.
ly
20
Protein Chem. 69: 73-100.
On
18
Sancar, A. (2004) Photolyase and cryptochrome blue-light photoreceptors. Adv
w
16
Res. 451: 25-37.
ie
15
Sancar, G. (2000) Enzymatic photoreactivation: 50 years and counting. Mutat
ev
14
rR
11
Singh, A., Doerks, T., Letunic, I., Raes, J., and Bork, P. (2009) Discovering
22
functional novelty in metagenomes: examples from light-mediated
23
processes. J Bacteriol. 191: 32-41.
24
25
Storey, J., and Tibshirani, R. (2003) Statistical significance for genomewide
studies. Proc. Natl. Acad. Sci. USA 100: 9440–9445
Wiley-Blackwell and Society for Applied Microbiology
27
Page 70 of 85
1
Strehl, B., Holtzendorff, J., Partensky, F., and Hess, W. (1999) A small and
2
compact genome in the marine cyanobacterium Prochlorococcus marinus
3
CCMP 1375: lack of an intron in the gene for tRNA(Leu)(UAA) and a
4
single copy of the rRNA operon. FEMS Microbiol Lett. 181: 261-266.
5
oceanic cyanobacterium Prochlorococcus. Nature 424: 1047-1051.
rP
Fo
6
Sullivan, M., Waterbury, J., and Chisholm, S. (2003) Cyanophages infecting the
7
Tedetti, M., and Sempere, R. (2006) Penetration of ultraviolet radiation in
8
the marine environment. a review. Photochem Photobiol. 82: 389-
9
397.
w
ie
ev
rR
ee
10
ly
On
Wiley-Blackwell and Society for Applied Microbiology
28
Page 71 of 85
1
Table 1. Primers used for qRT-PCR reactions
Gene
Primers
phrB
Forward: 5’-TGCAAGTTCAAGAGCTTGGTT-3’
Reverse: 5’-CGGATCTCCTTCTTCCAAAA-3’
Forward: 5’-AACTCTTGTCCCTGCTCCTG-3’
rP
Fo
mutT
Reverse: 5’-TCTCCTTTAACCTCGGAATTGA-3’
recA
Forward: 5’-CTTGGAGATGCCTCCAGAAT-3’
Reverse: 5’-CAACCACTCTTCCCTTTGGA-3’
ee
radA
Forward: 5’-GAAGGATCGAGACCATTTGC-3’
Reverse: 5’-GCTCATTCCAGTTGTTGTTCG-3’
rR
rnpB
Forward: 5’GAGGATAGTGCCACAGAAACATACC -3’
Reverse: 5’-GCTGGTGCGCTCTTACCACACCCTTG-3’
w
ie
ev
ly
On
Wiley-Blackwell and Society for Applied Microbiology
29
Page 72 of 85
Table 2. DNA sequence differences between Prochlorococcus MED4 strains
used in this study and the published parent MED4 genome (Rocap et al., 2003)
SNP(DNA)1
Amino Acid
GGA→AGA
G99R2
Translation initiation factor IF2
GGA→GGG
G367G3
Thioredoxin reductase
ATC →ATT
I63I3
N67N3
F202F3
Gene/Base Function
Recombination, DNA damage
recA
trxA
Cytochrome b subunit
AAT →AAC
metB
Amino acid transport/ metabolism
TTT →TTC
spoT
Guanosine polyphosphate
ev
rR
petD
ee
rP
Fo
infB
ie
AAT →AAG
pyrophosphohydrolase
N489K
w
Putative glycosyl transferase
CTA →CCA
L33P
1400750
Putative DNA helicase
AGT →ACT
S795T
208520
Putative sulfate transporter
GTG →TTG
V133L
1649373
Non-coding region
A →G
ly
On
1172708
Wiley-Blackwell and Society for Applied Microbiology
30
Page 73 of 85
1411655
Non-coding region
G→A
1052907
Non-coding region
X4 → G
940470
Non-coding region
X4 → A
683868
rP
Fo
683939
Non-coding region
G → X4
Non-coding region
G → X4
ee
Non-coding region
227758
Non-coding region
A → X4
T→A
ev
rR
276537
1
ie
Only the SNP in the shaded row is unique to MED4 UVR1, and thus represents
the single genomic sequence difference between the MED4 WT and MED4 UVR1
w
strains. The remaining SNPs were present in both MED4 WT and MED4 UVR1.
2
On
Position 99 corresponds to position 88 in the E. coli recA gene. This
unconserved residue is frequently G, but variants with R substitutions have no
mutations.
4
X denotes
absence of a base.
ly
known phenotypic consequences (Karlin and Brocchiere, 1996). 3 Denotes silent
Wiley-Blackwell and Society for Applied Microbiology
31
Page 74 of 85
Table 3. Sequences upstream of the mutT-phrB
operon in wild Prochlorococcus strains
Strain or GOS1 Read
Relevant sequence upstream of
number
mutT-phrB
rP
Fo
MED4 WT
AAAAAA
MED4 UVR1, UVR2, UVR3
AAAA -A
ee
JCVI_READ_1103359078035 AAAA -A
rR
JCVI_READ_1093017692266 AAAAGA
JCVI_READ_1093018630368 AAAAGA
ev
JCVI_READ_1103769464916 AAAAGA
JCVI_READ_1103769440186 AAAAGA
JCVI_READ_1103359190125 AAAAGA
JCVI_READ_ 843287
AAAAAA
JCVI_READ_1519837
AAAAAA
JCVI_READ_1568009
AAAAAA
JCVI_READ_1653138
AAAAAA
JCVI_READ_1664920
AAAAAA
ly
AAAAAA
On
JCVI_READ_ 630691
w
JCVI_READ_1103242864663 AAAAGA
ie
JCVI_READ_1103769744552 AAAAGA
Wiley-Blackwell and Society for Applied Microbiology
32
Page 75 of 85
JCVI_READ_1091141237359 AAAAAA
JCVI_READ_1092256175212 AAAAAA
JCVI_READ_1092963601548 AAAAAA
JCVI_READ_1093017692266 AAAAAA
JCVI_READ_1092344011058 AAAAAA
rP
Fo
JCVI_READ_1092347061561 AAAAAA
JCVI_READ_1092344012813 AAAAAA
JCVI_READ_1103769374878 AAAAAA
JCVI_READ_1102140312344 AAAAAA
ee
JCVI_READ_1105333561959 AAAAAA
JCVI_READ_1103668709694 AAAAAA
rR
JCVI_READ_1103242175527 AAAAAA
JCVI_READ_1105333435622 AAAAAA
ev
JCVI_READ_1104230100283 AAAAAA
JCVI_READ_1108840035938 AAAAAA
w
JCVI_READ_1108830123070 AAAAAA
ie
JCVI_READ_1103242847341 AAAAAA
1
http://camera.calit2.net/index.php
ly
On
The GOS database can be accessed at
Wiley-Blackwell and Society for Applied Microbiology
33
Page 76 of 85
Figure Legends
2
Figure 1. Enrichment selection for UV-hyper-resistant Prochlorococcus MED4
3
(MED4 UVR1). Cells from a mid-log phase culture of MED4 WT were subjected
4
to 30 seconds of UV exposure (254 nm) on Day 0, then returned to standard
5
growth conditions. This was done for 10 consecutive rounds of UV exposure and
6
re-growth (labeled 1 – 10). ○, rounds 1-5; ●, rounds 6-10.
7
rP
Fo
1
8
Figure 2. Sensitivity to UV light of MED4 WT and MED4 UVR1. MED4 WT (A, B,
9
C) and MED4 UVR1 (D, E, F) were treated with UV light at 3 wavelengths: 254,
ee
302, and 365 nm (UVC, UVB, and UVA, respectively), for various time intervals on
11
Day 0. Following UV treatment, cells were incubated under standard growth
12
conditions and monitored over time, using relative culture fluorescence as a
13
measure of cell density in the cultures. Values are the mean of measurements of
14
duplicate cultures, with error bars showing the standard deviation.
ie
ev
rR
10
15
Figure 3. Expression of phrB and mutT in MED4 UVR1 relative to MED4 WT
17
in log phase cells without UV treatment. Expression of phrB (encoding
18
photolyase) and mutT (encoding nudix hydrolase) was measured using whole
19
genome microarrays. These were the only genes that were significantly
20
differentially expressed between MED4 WT and MED4 UVR1. Ratios are the
21
mean values +/- standard error for 2 microarray experiments, each carried out in
22
duplicate.
w
16
ly
On
23
24
Figure 4. Expression of DNA repair genes determined by qRT-PCR in MED4
25
UVR1 and MED4 WT, with and without UV treatment. Ratios are the mean
Wiley-Blackwell and Society for Applied Microbiology
34
Page 77 of 85
values +/- standard error for 2 independent PCR runs, each carried out in
2
duplicate. A: Transcript levels of 4 genes in MED4 UVR1 relative to MED4 WT in
3
log phase cells in the absence of UV treatment. B and C: Transcript levels of the
4
same 4 genes in log phase cultures subjected to brief UV treatment (254 nm, 30
5
sec), relative to untreated controls. Transcript levels at 5 and 60 minutes after UV
6
treatment are shown for MED4 WT (B) and MED4 UVR1 (C).
7
rP
Fo
1
8
Figure 5. Sensitivity to UV of high light-adapted (MED4 WT) vs. low light-adapted
9
(MIT9313) strains of Prochlorococcus. Mid-log phase cells were treated with UV
light for various time intervals (A: UVC for 30 sec, B: UVB for 5 min, and C:
11
UVA,for 1 h) on Day 0. Following UV treatment, cells were incubated under
12
standard growth conditions and monitored for culture fluorescence over time.
13
Fluorescence of UV-treated cultures is expressed as per cent of untreated culture
14
fluorescence on Day 0. Values are the mean of measurements of duplicate
15
cultures, with error bars showing the standard deviation.●, MED4 WT; ○,
16
MIT9313.
w
ie
ev
rR
ee
10
17
On
Figure 6. The MED4 mutT-phrB operon and location of UV hyper-resistance
19
mutations in hyper-resistant mutants. The mutT-phrB operon of Prochlorococcus
20
MED4 WT is flanked by two unrelated structural genes, folK and degT
21
(http://www.microbesonline.org/cgibin/fetchLocus.cgi?locus=565622&disp=1).
22
The highlighted sequence shows the intergenic region between folk and mutT,
23
including the location of the homopolymer region (enlarged font) containing the
24
1 bp deletion (underlined) in the MED4 UVR strains as compared with MED4 WT.
25
The intergenic region of MED4 WT was sequenced twice, first as the parent of
ly
18
Wiley-Blackwell and Society for Applied Microbiology
35
Page 78 of 85
1
MED4 UVR1, and then verified a year later as the parent of MED4 UVR2 and
2
MED4 UVR3.
3
Supplementary Figure 1. Cell fluorescence following UV treatment is proportional
5
to cell number. Cells from mid-log phase cultures of MED4 WT and MED4 UVR1
6
were subjected to 30 seconds of UV exposure (254 nm) on Day 0, then returned
7
to standard growth conditions. Untreated (○) and UV-treated (●) cultures were
8
sampled for culture fluorescence (panels A and B), and for cell number using flow
9
cytometry (panels C and D). Values are the mean of measurements of duplicate
12
cultures, with error bars showing the standard deviation.
w
ie
ev
rR
11
ee
10
rP
Fo
4
ly
On
Wiley-Blackwell and Society for Applied Microbiology
36
Fo
Relative culture fluorescence
Page 79 of 85
1000
Selection for MED4 UVR1
8
rP
9
7
10
ee
5
100
3
4
2
rR
10
0
2
6
ev
4
1
iew
6
8
10
Time (days after UV treatment)
On
ly
Figure 1
Wiley-Blackwell and Society for Applied Microbiology
Page 80 of 85
MED4 WT
MED4 UVR1
1000
A.
D.
100
Fo
rP
10
ee
254 nm (UVC)
Relative culture fluorescence
no UV
15 seconds
30 seconds
60 seconds
1
1000
254 nm (UVC)
rR
B.
ev
100
1
1000 0
no UV
5 minutes
10 minutes
15 minutes
iew
10
302 nm (UVB)
C.
10
On
302 nm (UVB)
ly
F.
no UV
30 minutes
45 minutes
60 minutes
100
Figure 2
E.
365 nm (UVA)
365 nm (UVA)
0
2
4
6
8
10 0
2
4
Time (days after UV treatment)
Wiley-Blackwell and Society for Applied Microbiology
6
8
10
Page 81 of 85
Log2 (MED4 UVR1 / MED4 WT)
3.5
Fo
3.0
2.5
rP
2.0
1.5
1.0
0.5
0.0
ee
rR
ev
iew
mutT
phrB
(Photolyase) (Nudix
hydrolase)
On
ly
Figure 3
Wiley-Blackwell and Society for Applied Microbiology
Page 82 of 85
Fo
MED4 UVR1/ MED4 WT
No UV treatment
MED4 WT
ee
A.
8
6
4
2
0
-2
Log2 (UV treated / no UV treatment)
Log2 (MED4 UVR1 / MED4 WT)
10
rP
UV Treated / no UV treatment
8
B.
C.
rR
6
4
ev
iew
2
0
5’ 60’
-2
phrB recA
mutT radA
MED4 UVR1
phrB
5’ 60’
5’ 60’
minutes after UV
recA
mutT
On
5’ 60’
radA
Figure 4
Wiley-Blackwell and Society for Applied Microbiology
ly
5’ 60’
phrB
5’ 60’
5’ 60’
minutes after UV
recA
mutT
5’ 60’
radA
Page 83 of 85
Per cent of initial (untreated)
relative fluorescence
Fo
rP
120
A. UVC
MED4 WT
100
80
300
ee
250
200
60
150
40
100
20
500
B. UVB
400
rR
MED4 WT
iew
MIT9313
0
2
4
6
8
10
200
MED4 WT
100
On
0
0
0
MIT9313
300
ev
50
MIT9313
C. UVA
0
2
4
6
8
10
Time (days after UV treatment)
Figure 5
Wiley-Blackwell and Society for Applied Microbiology
0
2
ly
4
6
8
10
Page 84 of 85
Fo
rP
Putative regulatory region
DNA repair genes
Nudix hydrolase Photolyase
ee
folK
folk (3’ end)
phrB
mutT
degT
rR
mutT (5’ end)
ev
GTAGAAGCCAAAAAATCAAAAAGTTTGATGCCCCAAAATGGACTATATAAATGATTTACAAAACCAAAAAA TAACTGTTAGGCTGTCTTTATTTAAAAATTATG MED4 WT
GTAGAAGCCAAAAAATCAAAAAGTTTGATGCCCCAAAATGGACTATATAAATGATTTACAAAACCAAAAA
iew
TAACTGTTAGGCTGTCTTTATTTAAAAATTATG MED4 UVR1
GTAGAAGCCAAAAAATCAAAAAGTTTGATGCCCCAAAATGGACTATATAAATGATTTACAAAACCAAAAA
TAACTGTTAGGCTGTCTTTATTTAAAAATTATG MED4 UVR2
GTAGAAGCCAAAAAATCAAAAAGTTTGATGCCCCAAAATGGACTATATAAATGATTTACAAAACCAAAAA
TAACTGTTAGGCTGTCTTTATTTAAAAATTATG MED4 UVR3
On
Intergenic region between folk and mutT
ly
Figure 6
Wiley-Blackwell and Society for Applied Microbiology
Page 85 of 85
Relative culture fluorescence
1000
1000
MED4
MED4 UVR1
A. MED4
100
B. MED4 UVR1
Fo
100
rP
0
2
4
ee
6
8
rR
10
0
ev
1e+9
MED4
C. MED4
2
MED4
UVR1
D. MED4
Cells ml-1
1e+8
0
2
4
6
8
10
6
8
10
1e+9
iew
1e+8
4
UVR1
On
ly
0
2
4
Time (days after UV treatment)
Supplementary Fig. 1
Wiley-Blackwell and Society for Applied Microbiology
6
8
10