AEM Accepts, published online ahead of print on 26 February 2010 Appl. Environ. Microbiol. doi:10.1128/AEM.01992-09 Copyright © 2010, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved.
1
Metabolomic characterization of the salt stress response in Streptomyces
2
coelicolor
3 4
Stefan Kol1+, M. Elena Merlo1,2+, Richard A. Scheltema2+, Marcel de Vries3, Roel J.
5
Vonk3, Niels A. Kikkert2, Lubbert Dijkhuizen1, Rainer Breitling2*, and Eriko Takano1*
6 7
1
8
University of Groningen, Kerklaan 30, 9751 NN Haren, The Netherlands
9
2
Microbial Physiology, Groningen Biomolecular Sciences and Biotechnology Institute,
Groningen Bioinformatics Centre, Groningen Biomolecular Sciences and Biotechnology
10
Institute, University of Groningen, Kerklaan 30, 9751 NN Haren, The Netherlands
11
3
12
Groningen, The Netherlands
Centre for Medical Biomics, University Medical Centre Groningen, 9713 AV
13 14
*
15
+31-50-3632154 (Streptomyces physiology) or R. Breitling,
[email protected], Tel. +31-
16
50-3638088 Fax. +31-50-3637976 (metabolomics)
17
+
Address correspondence to: E. Takano,
[email protected], Tel. +31-50-3632143 Fax.
Equally contributing authors
18 19
Journal section: physiology
20
Running title: Metabolomics of S. coelicolor
21
1
22
Abbreviations used: LC-MS, liquid chromatography mass spectrometry; ASA, L-
23
aspartate-β-semialdehyde; DABA, L-2,4-diaminobutyrate; ADABA, N-γ-acetyl-L-2,4-
24
diaminobutyrate, Ect, ectoine; EctOH, hydroxyectoine; HILIC, hydrophobic interaction
25
liquid chromatography.
26 27
Keywords: osmoadaptation, metabolomics, LTQ-Orbitrap, ectoine, salt stress,
28
Streptomyces coelicolor
2
29
Abstract
30
The humicolous actinomycete, Streptomyces coelicolor, routinely adapts to a wide
31
variety of habitats and rapidly changing environments. Upon salt stress the organism is
32
also known to increase the levels of various compatible solutes. Here we report the
33
results of the first high resolution metabolomics time series analysis on various strains of
34
S. coelicolor exposed to salt-stress: wild-type, progressive knockouts of the ectoine
35
biosynthesis pathway, and two stress regulator mutants (disruptions of the sigB and osaB
36
genes). Samples were taken from cultures at 0, 4, 8 and 24 h after salt-stress treatment
37
and analyzed by liquid chromatography mass spectrometry on the LTQ-Orbitrap XL. The
38
results suggest that a large fraction of amino acids is upregulated in response to the salt
39
stress, as well as proline/glycine-containing di- and tripeptides. Additionally we found
40
that 5’-methylthioadenosine, a known inhibitor of polyamine biosynthesis, is
41
downregulated upon salt stress. Strikingly, no major differences between the wild-type
42
cultures and the two stress regulator mutants were found, indicating considerable
43
robustness of the metabolomic response to salt stress, compared to the more volatile
44
changes in transcript abundance reported earlier.
45
[180 / 250]
3
46
Introduction
47
Salt stress conditions are known to cause major changes in the primary and secondary
48
metabolism of bacterial cells (26, 34). This is also the case for the Gram-positive soil
49
bacteria of the genus Streptomyces, which are well-known for their complex secondary
50
metabolism and production of a wide range of industrially relevant metabolites, including
51
various antibiotics (21). Upon salt stress, these species are known to increase the levels of
52
various compatible solutes, including ectoine, alanine, glutamine and proline (6, 11, 16,
53
28). The genome sequence of Streptomyces indicates a largely unexplored capacity for
54
secondary metabolite production (2), some of which could also be involved in the salt
55
stress response. Here we conducted an untargeted metabolomics analysis of salt-stressed
56
Streptomyces coelicolor, the genome-sequenced model species of the genus (2), to
57
investigate the complexity of the metabolite changes associated with the salt stress
58
response in more detail and to increase our understanding of the mechanisms of
59
osmoadaptation by metabolomic rearrangement. For this purpose we used a high
60
resolution
61
chromatography (23). The accurate mass determination and high resolving power of the
62
Orbitrap (22, 40) have made the analysis of complex metabolite mixtures feasible,
63
making it an attractive option for untargeted metabolomics screens (5). The use of HILIC
64
liquid chromatography ensures that the resolving power of the mass spectrometer is
65
further extended by fractionation of the sample, resulting in a less complex mixture being
66
injected in the mass spectrometer at each moment in time. In addition to wild type S.
67
coelicolor we also studied strains disrupted in various steps of the biosynthesis of ectoine
68
and hydroxyectoine (Figure 1), the two best-characterized osmoprotectants of
LTQ-Orbitrap
mass
spectrometer
(MS)
coupled
to
HILIC
liquid
4
69
Streptomyces (6, 7), as well as mutants of two stress regulators (OsaB and σB) under
70
continuous salt stress and during a time course following salt shock. The two stress
71
regulators were previously shown to be involved in the salt shock response (3, 32). This
72
study provides the first attempt at a global metabolomic characterization of the salt
73
adaptation process in bacteria.
74
Materials and methods
75
Strains and media conditions
76
Escherichia coli strain ET12567/pUZ8002 was used for conjugation of cosmids into S.
77
coelicolor (14). Salt shock or osmoadaptation experiments using S. coelicolor M145
78
parent strain (27), LW35 (ectA::Tn) (this study), LW36 (ectC::Tn) (this study), LW37
79
(ectD::Tn) (this study), osaB::Tn (3) and sigB::aac(3)IV (∆sigB) (31) were performed in
80
Difco Nutrient Broth (DNB), whereas solid growth experiments were performed on
81
Supplemented Minimal Media (SMMS) supplemented with the indicated amounts of
82
NaCl or ectoines (20). For the liquid cultures, a high inoculum of spores was used to
83
reduce the formation of clumps.
84
Ectoine mutant construction and characterization
85
Three cosmids containing Tn5062 transposon insertions in ectA (SCO1864), ectC
86
(SCO1866) and ectD (SCO1867) were used to disrupt the ectoine biosynthesis cluster
87
genes in S. coelicolor. Conjugation to S. coelicolor was performed as previously reported
88
(3). Allelic replacements were screened by selecting for AprR and KanS clones. To
89
confirm the genomic replacement of the ectA, ectC, and ectD genes by the Tn5062-
90
containing copies, genomic DNA from each mutant was isolated and used as a template 5
91
for PCR using either the transposon primers EZR1 and EZL2 (3) and ect gene primers
92
which anneal to the start and stop codons of ectA, ectC and ectD. PCR conducted with
93
wild type strain M145 total DNA as a template gave rise to amplified products which
94
corresponded in size with the predicted products of 633 (ectA), 1250 (ectB), 414 (ectC)
95
and 888 (ectD) basepairs using the ect gene primers (Supplementary Figure A.b, lane 1–
96
4). Due to the insertion of the large transposon, no PCR product was amplified when only
97
the ect gene primers were used on genomic DNA from the mutants (Supplementary
98
Figure A.b, lane 7, 10 and 13). However, the transposon primers in combination with the
99
ect gene primers yielded products corresponding in size with the predicted distance from
100
the ect start and stop codons to the location of the transposon (Supplementary Figure A.b,
101
lane 5, 6, 8, 9, 11, 12). These data confirmed the genomic disruption of the ectA, ectC,
102
and ectD genes. This was further validated by a complementation analysis using a drop
103
dilution assay. Cells were grown on SMMS with 1M NaCl, with and without additional
104
20µM ectoine and hydroxyectoine, and monitored for growth (Figure 2).
105
Growth conditions and metabolite extraction
106
Flasks containing 50 ml DNB were inoculated with 1·108 S. coelicolor spores of M145,
107
ectA::Tn, ectC::Tn, ectD::Tn, osaB::Tn and ∆sigB and grown for 24 hours, at which
108
point 5 ml of 5 M NaCl or 5 ml of MilliQ water was added for salt shock samples. Cells
109
exposed to continuous salt stress were inoculated and grown in 50 ml DNB supplemented
110
with 0.5 M NaCl. Samples of 5 ml (OD450 of 1.0), or an equivalent thereof to obtain
111
identical cell mass, were harvested at 0, 4, 8 and 24 hours after addition of salt for the salt
112
shock cultures or at 24 and 48 hours for the continuously salt-stressed cultures. Cells
113
were collected by centrifugation for 10 minutes at 4500 rpm and 4°C, the supernatant was 6
114
removed completely and the cell pellet was resuspended in 500 µl methanol. After
115
incubation on ice for 10 min with occasional mixing, cells were spun down at 4°C, 13000
116
rpm, and the supernatant was collected and stored at –80°C.
117
Liquid Chromatography Mass Spectrometry
118
Data were acquired with a Luna 3µ HILIC 200A HPLC (150 x 2.0 mm; Bester,
119
Rotterdam, The Netherlands), coupled to an LTQ-Orbitrap XL (Thermo Fisher Scientific,
120
Bremen, Germany) in positive ionization mode at a resolution of 30.000. The LC-MS
121
system was run in binary gradient mode with a flow-rate of 400 µl/min; a volume of 20 µl
122
methanol samples was injected. The solvents used were A: 90/10 acetonitrile with 5 mM
123
ammonium acetate and B: 50/50 acetonitrile with 5 mM ammonium acetate. The used
124
gradient was set to hold 90% acetonitrile isocratic for 2.5 minutes, after which it was
125
moved to 50% in 7.5 minutes, and finally held isocratic for 2.5 minutes. The same
126
conditions were maintained for the MS/MS measurements. Analysis of the fragmentation
127
patterns was done in XCalibur 2.0.5. Extracts were supplemented before injection with
128
solvent A in a ratio of 1:10, improving the LC peak shape. Even though molecules were
129
diluted 10-fold, the peak intensity levels of compounds of interest were not affected
130
negatively due to sharper peaks. Ectoine and hydroxyectoine standards (Fluka,
131
Zwijndrecht) were used as controls.
132
Data processing
133
Data processing was performed using a configurable software pipeline, capable of
134
extracting and automatically matching peaks from multiple measurements into peak
135
groups (40). Retention time alignment of the extracted mass chromatograms was done
7
136
using the COW-CoDA algorithm (9). Noise resulting from the greedy peak picking
137
approach was reduced by application of an RSD (relative standard deviation) filter set to
138
35% for matched replicates, as quantification is expected to be at least 20% accurate over
139
multiple runs (41). The selected mass chromatograms were putatively identified by
140
matching the masses progressively to metabolite specific databases. First ScoCyc (2),
141
LipidMAPS (12) and a contaminant database (25) were used. Unidentified peak groups
142
were then matched to KEGG (24) and the remaining unidentified to Metlin (42) and the
143
Human Metabolome Database (47). This iterative process was used in order to restrict the
144
number of potential matches to the most likely (39). As a last step all peak groups related
145
to another peak group (e.g., isotopes, ion adducts, and fragments) were removed from the
146
set based on a correlation analysis of the intensity patterns and peak shape (46).
147
Hierarchical clustering was performed to identify metabolites that showed similar
148
dynamic changes after salt shock in various mutants. To achieve a robust clustering, the
149
distance score was calculated for data with sufficient variation (RSD > 0.2) based on the
150
Pearson correlation of the discretized slopes of the z-score normalized time courses,
151
rather than on the absolute values. By this approach, instead of determining if the original
152
signal intensities correlate, we determined the correlation of the slopes, i.e. the pattern of
153
changes over time, and in order to reduce the influence of noise on the slopes, they were
154
discretized to three possible values (–1, 0, or 1) depending on whether signal intensities
155
were decreasing, stable, or increasing. The resulting hierarchical tree was consequently
156
cut into 4 clusters. For each cluster a quality score was calculated, based on the average
157
distance of each observation of each strain to the average profile of the cluster; this
158
values indicates the coherence of expression profiles within the cluster (smaller values
8
159
indicate better quality). For convenient visualization of the time courses of individual
160
metabolites, the intensity values are displayed as fold-changes relative to the initial time
161
point (i.e., for each strain they are divided by the intensity value of the first time point of
162
that strain).
163
Results and discussion
164
Validation of metabolite quantification
165
The relative quantification accuracy of the mass spectrometry data was assessed by
166
quantifying the S. coelicolor antibiotic undecylprodigiosins, which are visibly present in
167
the extracts (the extracts are dark red when high amounts of antibiotics are produced).
168
The traditional method of measuring the abundance of undecylprodigiosins by optical
169
density (OD) measurement at 533nm (44) was used as a second quantification method.
170
Supplementary Figure B shows the total detected signal for undecylprodigiosin (mass
171
391.26236: butylcycloheptylprodigiosin, metacycloprodiginine and methyl-cyclo-decyl-
172
prodiginine; 393.27801: undecylprodiginine) compared to the OD533nm values. The
173
Pearson correlation is 0.95, indicating that the relative quantification achieved by the
174
HILIC-Orbitrap combination is surprisingly precise. Especially considering that these
175
compounds do not bind well to the HILIC column and are flushed out with a large
176
number of other compounds (mostly lipids), making exact quantification challenging due
177
to ionization suppression effects (23). For the large number of more polar compounds in
178
our metabolome screen, which are better separated by the HILIC column, we can
179
therefore expect similarly reliable quantitation.
9
180
Metabolomics of continuous salt exposure
181
To validate our methodology, we first focused on a targeted metabolomics screen of wild
182
type and ectoine biosynthesis mutants of S. coelicolor grown under continuous salt stress.
183
Under these conditions, we expect very characteristic abundance profiles for ectoine,
184
hydroxyectoine and their precursors; these critical osmoprotectants should be absent in
185
the mutants that show a clear growth defect under salt-stress conditions (Figure 2). This
186
is indeed the case in our measurements, as the found intensity levels closely matched the
187
expected phenotype of the knockouts (Figure 3): (1) The early precursor metabolite
188
DABA was found to be accumulating in the ectA::Tn strain, where its conversion to
189
ADABA is deficient, and was detected nowhere else. (2) The later precursor metabolite
190
ADABA was found to accumulate strongly in the ectC::Tn strain and, to a lesser degree,
191
in the ectD::Tn strain. The presence of this metabolite in the ectD::Tn strain can be
192
explained by the reversibility of its conversion to ectoine. (3) Ectoine was found to
193
accumulate strongly in the ectD::Tn strain, where its conversion to hydroxyectoine is
194
deficient. Low intensity levels of ectoine were also detected for the M145 strain,
195
suggesting that ectoine acts as an intermediate in S. coelicolor and is converted almost
196
quantitatively to hydroxyectoine. Very low levels of the same mass and similar retention
197
time as ectoine were also detected for the ectA::Tn and ectC::Tn strains; this could
198
indicate the existence of an alternative, low-yield biosynthesis pathway for ectoine or the
199
presence of an unrelated isomer of low abundance. In both cases, the detected levels were
200
so low that we do not expect this to be of biological relevance for osmoprotection. (4)
201
Finally, hydroxyectoine was found to accumulate strongly in the M145 strain, indicating
202
that it is acting as the prime osmoprotectant. Retention time and tandem MS
10
203
fragmentation patterns in comparison with the standards confirmed the identity of the
204
detected compounds as ectoine and hydroxyectoine in the strains ectD::Tn and M145
205
respectively (Supplementary Figure C and D).
206
Metabolomics of the salt shock response
207
Global metabolome screen
208
To obtain a global metabolomic characterization of the salt shock response, we measured
209
time series of metabolite profiles in salt shocked cells, using the same S. coelicolor parent
210
(M145) and mutant strains (ectA::Tn, ectC::Tn, ectD::Tn, osaB::Tn and ∆sigB) as before.
211
1247 distinct peak groups were quantified, which were combined into 363 peak groups of
212
related peaks (each corresponding to a potential metabolite). Of the peak groups, 229
213
could be assigned putative identities based on exact mass matching (for an overview of
214
the identifications see Figure 4 and supplementary Table A). This compares well to the
215
predicted amount of metabolites based on genome annotations reported by Borodina et al
216
(4).
217
Dynamic responses to salt exposure
218
The dynamic response to salt exposure was visualized by unsupervised clustering of the
219
metabolite levels for the time series of the salt-stressed cultures. Two coherent clusters of
220
metabolites were revealed with consistently increasing and decreasing abundance (Figure
221
5, cluster 1 and 4 respectively). All other clusters are much less coherent as indicated by
222
their quality scores. This coherence shows that the global metabolomic response to salt
223
stress is clearly dominating in our samples and involves a reasonably large group of 52
224
peak groups (33 increasing; 19 decreasing). Comparison of the time trends of all 11
225
putatively identified metabolites from the salt shock to the control non-salt shocked
226
cultures revealed reproducible results for 15 metabolites accumulating in response to the
227
salt shock, and 3 metabolites with the reverse response (Table 1). The remaining
228
metabolites either did not give reproducible results or showed equal behavior in the salt
229
shock and control cultures. Figure 6 shows exemplary time courses with and without salt
230
shock, including the putative molecular identities of the compounds involved. A large
231
part of the upregulated compounds are also contained in the medium used and further
232
analysis is required to determine whether they are synthesized or taken up by the
233
organism. The most important accumulating metabolite is proline, which is a well-known
234
osmoprotectant that was previously reported for Streptomyces (11). The strong
235
accumulation of proline after the salt shock is the same in all the strains, while in the non-
236
shocked cultures proline levels remain stable or slightly decrease over time. The strong
237
and immediate change indicates that proline is used for acute osmoprotection.
238
Interestingly, this response is independent of the major regulators OsaB and σB. In
239
addition to this classical salt-responsive metabolite, the global metabolome screen also
240
identified additional potential osmoprotectants. Arginine, phenylalanine, methionine,
241
tryprophan and iso-/leucine show the same strong accumulation as proline, and thus
242
could also play a supplementary role in acute osmoprotection. Like all amino acids, the
243
listed compounds are zwitterions that could plausibly act as osmoprotectants. A similar
244
dynamic, seen for guanosin, adenosine, adenine, and hypoxanthine, nucleotide
245
derivatives, is less obvious to explain. While affecting a large number of amino acids, the
246
accumulation in response to the salt shock is not a general phenomenon for all amino
247
acids, as can be seen in the time courses for glutamate and valine.
12
248
In addition to the amino acids, we detected five di- and tripeptides that also
249
accumulated during salt stress (Figure 4). All of these peptides contained a proline and a
250
glycine residue. These proline/glycine containing peptides were significantly enriched
251
compared to the protein composition of both S. coelicolor and Saccharomyces cerevisiae
252
(p-value < 0.01, based on random sampling of peptide sequences). Glycine, glutamine
253
and alanine individually were observed at similar significance levels, while other amino
254
acids were represented at the same frequency in proteins and the accumulating peptides.
255
This indicates that proline/glycine-containing peptides are either produced by specific
256
proteolysis or taken up preferentially from the peptone containing medium. A possible
257
candidate for a di- and tripeptide uptake system is SCO3064, which has homology with
258
DtpT from Lactococcus lactis (18). This is in agreement with earlier studies in other
259
organisms that also showed an involvement of proline-containing peptides in the salt
260
stress response (1, 30, 37).
261
One of the few putatively identified metabolites that showed a slight, but
262
consistent decrease in abundance in salt shock, but not in control conditions was 5’-
263
methylthioadenosine. This metabolite acts as an inhibitor of polyamine biosynthesis in
264
vitro and in vivo (19, 36). As polyamines are known to protect against salt stress in plants
265
(48), an active degradation of the inhibitor 5’-methylthioadenosine could be part of the
266
salt response in Streptomyces as well. The pfs gene encoding 5'-methylthioadenosine/S-
267
adenosylhomocysteine nucleosidase (10) is increased in expression by salt stress in E.
268
coli (29). Alternatively, the decreased levels of 5’-methylthioadenosine could indicate a
269
decreased activity of the polyamine biosynthesis pathway under salt stress.
13
270
Distinguishing these two alternatives will require targeted measurements of polyamines,
271
which were not detected in our experimental setup.
272
Other metabolites that decreased in abundance in high-salt conditions also did so
273
in the control cultures. For instance, the amino acid histidine and its catabolite urocanate
274
decrease in both cultures, the only amino acid-related compounds which show this
275
behavior.
276
Not surprisingly, ectoine and hydroxyectoine are absent from the core salt shock
277
response cluster, due to the disruption of the ectoine pathway in several of the strains.
278
The detected compounds are thus those that show a consistent immediate response to salt
279
shock independent of genotype, indicating that the salt shock metabolome extends far
280
beyond the ectoines. The only putative osmoprotectant that we detected, but which
281
showed no response to salt stress was alanine, which was previously described as an
282
osmoprotectant for Streptomyces (11). The time-series for all strains show stable behavior
283
of this metabolite for the salt shocked samples and the non-salt shocked control cultures,
284
indicating that it is in fact not used as osmoprotectant in the conditions tested.
285
Conclusions
286
Streptomyces species are known to be able to withstand considerable levels of osmotic
287
stress (34) and to accumulate ectoine and hydroxyectoine under salt adaptation (7). We
288
have studied the global metabolic response of S. coelicolor to salt stress, using the high
289
resolution LTQ-Orbitrap mass spectrometer. We did not aim at a complete coverage of
290
the entire metabolome, but rather intended to achieve an unbiased, comprehensive
291
assessment of the moderately polar osmolytes most similar to ectoine, in particular amino
292
acids, nucleotides and their derivatives. Sugars and polyols, the remaining major group of 14
293
osmoprotectants which were not detectable with the liquid chromatography mass
294
spectrometry conditions used, will therefore be the target of future studies using
295
complementary analytical technologies with higher selectivity towards these compounds.
296
It is very likely that among the described compounds, we have detected all that are
297
osmotically relevant, as osmoprotectants need to accumulate in large amounts to be
298
effective. However, the mass spectrometry data do not provide absolute quantification of
299
the compounds involved. Instead, our results are focusing on the dynamic changes of the
300
metabolome. This relative quantification, which is standard practice in metabolomics
301
studies, clearly identifies the temporal trends in a large number of potentially
302
osmoprotective compounds.
303
The current study leads to the putative identification of potential novel
304
osmoprotectants in S. coelicolor. Strikingly, we do not observe a major difference
305
between the metabolomic response to salt stress in wild-type cultures and the two
306
mutants in the major osmotic stress regulatory genes, osaB and sigB (32). This indicates a
307
considerable robustness of the metabolome, compared to the much more volatile
308
transcriptome, in agreement with recent studies in plants (15). Robustness of metabolic
309
fluxes under environmental and genetic perturbations had been found earlier (13, 45), and
310
seem to be a general feature of microbial metabolic networks (43). Our findings extend
311
this observation to the robustness of steady-state metabolite levels. This robustness could
312
be due to redundancy in the regulatory input leading to metabolic adaptation and has
313
obvious evolutionary benefits. Elucidating the cellular networks that enable this
314
metabolic robustness will require additional experimentation at multiple molecular levels,
15
315
including transcriptomics, proteomics and, most importantly, the salt-dependent kinetics
316
of the enzymes involved.
317
Furthermore, the strong accumulation of ectoine in the EctD::Tn strain is an
318
indication that S. coelicolor could potentially be used as a producer of ectoine in (large
319
scale-) bacterial milking (38), but the absolute levels of production would need to be
320
quantified to determine whether this is an attractive option.
321
322
Acknowledgements
323
The authors would like to thank Prof. Dr. Erhard Bremer for his critical reading of the
324
manuscript. The work was supported by grants from: EU-FP6 ActinoGEN (LD, SK), the
325
Northern Netherlands Collaboration Initiative (SNN EZ/KOMPAS RM119; LD), the
326
University of Groningen (Ubbo Emmius Fellowship, MEM; Rosalind Franklin
327
Fellowship, ET), and the Netherlands Organization for Scientific Research NWO (Vidi
328
grant, RB; medium investment grant, RJV).
16
329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374
References 1.
2.
3.
4. 5. 6.
7.
8.
9.
10.
11. 12. 13.
Amezaga, M. R., I. Davidson, D. McLaggan, A. Verheul, T. Abee, and I. R. Booth. 1995. The role of peptide metabolism in the growth of Listeria monocytogenes ATCC 23074 at high osmolarity. Microbiology 141 ( Pt 1):41-9. Bentley, S. D., K. F. Chater, A. M. Cerdeno-Tarraga, G. L. Challis, N. R. Thomson, K. D. James, D. E. Harris, M. A. Quail, H. Kieser, D. Harper, A. Bateman, S. Brown, G. Chandra, C. W. Chen, M. Collins, A. Cronin, A. Fraser, A. Goble, J. Hidalgo, T. Hornsby, S. Howarth, C. H. Huang, T. Kieser, L. Larke, L. Murphy, K. Oliver, S. O'Neil, E. Rabbinowitsch, M. A. Rajandream, K. Rutherford, S. Rutter, K. Seeger, D. Saunders, S. Sharp, R. Squares, S. Squares, K. Taylor, T. Warren, A. Wietzorrek, J. Woodward, B. G. Barrell, J. Parkhill, and D. A. Hopwood. 2002. Complete genome sequence of the model actinomycete Streptomyces coelicolor A3(2). Nature 417:141-7. Bishop, A., S. Fielding, P. Dyson, and P. Herron. 2004. Systematic insertional mutagenesis of a streptomycete genome: a link between osmoadaptation and antibiotic production. Genome Res. 14:893-900. Borodina, I., P. Krabben, and J. Nielsen. 2005. Genome-scale analysis of Streptomyces coelicolor A3(2) metabolism. Genome Res. 15:820-9. Breitling, R., A. R. Pitt, and M. P. Barrett. 2006. Precision mapping of the metabolome. Trends Biotechnol. 24:543-8. Bursy, J., A. U. Kuhlmann, M. Pittelkow, H. Hartmann, M. Jebbar, A. J. Pierik, and E. Bremer. 2008. Synthesis and uptake of the compatible solutes ectoine and 5-hydroxyectoine by Streptomyces coelicolor A3(2) in response to salt and heat stresses. Appl. Environ. Microbiol. 74:7286-96. Bursy, J., A. J. Pierik, N. Pica, and E. Bremer. 2007. Osmotically induced synthesis of the compatible solute hydroxyectoine is mediated by an evolutionarily conserved ectoine hydroxylase. J. Biol. Chem. 282:31147-55. Canovas, D., C. Vargas, M. I. Calderon, A. Ventosa, and J. J. Nieto. 1998. Characterization of the genes for the biosynthesis of the compatible solute ectoine in the moderately halophilic bacterium Halomonas elongata DSM 3043. Syst. Appl. Microbiol. 21:487-97. Christin, C., A. K. Smilde, H. C. Hoefsloot, F. Suits, R. Bischoff, and P. L. Horvatovich. 2008. Optimized time alignment algorithm for LC-MS data: correlation optimized warping using component detection algorithm-selected mass chromatograms. Anal. Chem. 80:7012-21. Cornell, K. A., and M. K. Riscoe. 1998. Cloning and expression of Escherichia coli 5'-methylthioadenosine/S-adenosylhomocysteine nucleosidase: identification of the pfs gene product. Biochim. Biophys. Acta 1396:8-14. Empadinhas, N., and M. S. da Costa. 2008. Osmoadaptation mechanisms in prokaryotes: distribution of compatible solutes. Int. Microbiol. 11:151-61. Fahy, E., M. Sud, D. Cotter, and S. Subramaniam. 2007. LIPID MAPS online tools for lipid research. Nucleic Acids Res. 35:W606-12. Fischer, E., and U. Sauer. 2005. Large-scale in vivo flux analysis shows rigidity and suboptimal performance of Bacillus subtilis metabolism. Nat. Genet. 37:63640. 17
375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420
14.
15.
16. 17.
18.
19.
20. 21. 22. 23.
24. 25.
26.
27. 28.
29.
Flett, F., V. Mersinias, and C. P. Smith. 1997. High efficiency intergeneric conjugal transfer of plasmid DNA from Escherichia coli to methyl DNArestricting streptomycetes. FEMS Microbiol. Lett. 155:223-9. Fu, J., J. J. Keurentjes, H. Bouwmeester, T. America, F. W. Verstappen, J. L. Ward, M. H. Beale, R. C. de Vos, M. Dijkstra, R. A. Scheltema, F. Johannes, M. Koornneef, D. Vreugdenhil, R. Breitling, and R. C. Jansen. 2009. Systemwide molecular evidence for phenotypic buffering in Arabidopsis. Nat Genet 41:166-7. Galinski, E. A. 1995. Osmoadaptation in bacteria. Adv. Microb. Physiol. 37:272328. Garcia-Estepa, R., M. Argandona, M. Reina-Bueno, N. Capote, F. IglesiasGuerra, J. J. Nieto, and C. Vargas. 2006. The ectD gene, which is involved in the synthesis of the compatible solute hydroxyectoine, is essential for thermoprotection of the halophilic bacterium Chromohalobacter salexigens. J. Bacteriol. 188:3774-84. Hagting, A., J. vd Velde, B. Poolman, and W. N. Konings. 1997. Membrane topology of the di- and tripeptide transport protein of Lactococcus lactis. Biochemistry 36:6777-85. Hibasami, H., and A. E. Pegg. 1978. Differential inhibition of mammalian aminopropyltransferase activities. Biochem. Biophys. Res. Commun. 81:1398405. Hopwood, D. A. 1967. Genetic analysis and genome structure in Streptomyces coelicolor. Bacteriol Rev. 31:373-403. Hopwood, D. A. 2007. Streptomyces in Nature and Medicine: The Antibiotic Makers. Oxford University Press. Hu, Q., R. J. Noll, H. Li, A. Makarov, M. Hardman, and R. Graham Cooks. 2005. The Orbitrap: a new mass spectrometer. J. Mass Spectrom. 40:430-43. Kamleh, M. A., J. A. Dow, and D. G. Watson. 2009. Applications of mass spectrometry in metabolomic studies of animal model and invertebrate systems. Brief. Funct. Genomic. Proteomic. 8:28-48. Kanehisa, M., and S. Goto. 2000. KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 28:27-30. Keller, B. O., J. Sui, A. B. Young, and R. M. Whittal. 2008. Interferences and contaminants encountered in modern mass spectrometry. Anal. Chim. Acta. 627:71-81. Kempf, B., and E. Bremer. 1998. Uptake and synthesis of compatible solutes as microbial stress responses to high-osmolality environments. Arch. Microbiol. 170:319-30. Kieser, T., M. J. Bibb, M. J. Butter, K. F. Chater, and D. A. Hopwood. 2000. Practical Streptomyces genetics. The John Innes Foundation, Norwich. Killham, K., and M. K. Firestone. 1984. Salt Stress Control of Intracellular Solutes in Streptomycetes Indigenous to Saline Soils. Appl. Environ. Microbiol. 47:301-306. Kim, Y., C. M. Lew, and J. D. Gralla. 2006. Escherichia coli pfs transcription: regulation and proposed roles in autoinducer-2 synthesis and purine excretion. J. Bacteriol. 188:7457-63. 18
421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465
30.
31.
32.
33.
34.
35.
36.
37.
38. 39.
40.
41.
42.
43.
Le Marrec, C., E. Bon, and A. Lonvaud-Funel. 2007. Tolerance to high osmolality of the lactic acid bacterium Oenococcus oeni and identification of potential osmoprotectants. Int. J. Food Microbiol. 115:335-42. Lee, E. J., Y. H. Cho, H. S. Kim, B. E. Ahn, and J. H. Roe. 2004. Regulation of sigmaB by an anti- and an anti-anti-sigma factor in Streptomyces coelicolor in response to osmotic stress. J. Bacteriol. 186:8490-8. Lee, E. J., N. Karoonuthaisiri, H. S. Kim, J. H. Park, C. J. Cha, C. M. Kao, and J. H. Roe. 2005. A master regulator sigmaB governs osmotic and oxidative response as well as differentiation via a network of sigma factors in Streptomyces coelicolor. Mol. Microbiol. 57:1252-64. Louis, P., and E. A. Galinski. 1997. Characterization of genes for the biosynthesis of the compatible solute ectoine from Marinococcus halophilus and osmoregulated expression in Escherichia coli. Microbiology 143 ( Pt 4):1141-9. Malin, G., and A. Lapidot. 1996. Induction of synthesis of tetrahydropyrimidine derivatives in Streptomyces strains and their effect on Escherichia coli in response to osmotic and heat stress. J. Bacteriol. 178:385-95. Ono, H., K. Sawada, N. Khunajakr, T. Tao, M. Yamamoto, M. Hiramoto, A. Shinmyo, M. Takano, and Y. Murooka. 1999. Characterization of biosynthetic enzymes for ectoine as a compatible solute in a moderately halophilic eubacterium, Halomonas elongata. J. Bacteriol. 181:91-9. Pegg, A. E., R. T. Borchardt, and J. K. Coward. 1981. Effects of inhibitors of spermidine and spermine synthesis on polyamine concentrations and growth of transformed mouse fibroblasts. Biochem. J. 194:79-89. Piuri, M., C. Sanchez-Rivas, and S. M. Ruzal. 2003. Adaptation to high salt in Lactobacillus: role of peptides and proteolytic enzymes. J. Appl. Microbiol. 95:372-9. Sauer, T., and E. A. Galinski. 1998. Bacterial milking: A novel bioprocess for production of compatible solutes. Biotechnol. Bioeng. 57:306-13. Scalbert, A., L. Brennan, O. Fiehn, T. Hankemeier, B. Kristal, B. van Ommen, E. Pujos-Guillot, E. Verheij, D. Wishart, and S. Wopereis. 2009. Mass-spectrometry-based metabolomics: limitations and recommendations for future progress with particular focus on nutrition research. Metabolomics. Scheltema, R. A., A. Kamleh, D. Wildridge, C. Ebikeme, D. G. Watson, M. P. Barrett, R. C. Jansen, and R. Breitling. 2008. Increasing the mass accuracy of high-resolution LC-MS data using background ions: a case study on the LTQOrbitrap. Proteomics 8:4647-56. Shah, V. P., K. K. Midha, J. W. Findlay, H. M. Hill, J. D. Hulse, I. J. McGilveray, G. McKay, K. J. Miller, R. N. Patnaik, M. L. Powell, A. Tonelli, C. T. Viswanathan, and A. Yacobi. 2000. Bioanalytical method validation--a revisit with a decade of progress. Pharm. Res. 17:1551-7. Smith, C. A., G. O'Maille, E. J. Want, C. Qin, S. A. Trauger, T. R. Brandon, D. E. Custodio, R. Abagyan, and G. Siuzdak. 2005. METLIN: a metabolite mass spectral database. Ther. Drug Monit. 27:747-51. Stelling, J., U. Sauer, Z. Szallasi, F. J. Doyle, 3rd, and J. Doyle. 2004. Robustness of cellular functions. Cell 118:675-85.
19
466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486
44. 45.
46. 47.
48.
Strauch, E., E. Takano, H. A. Baylis, and M. J. Bibb. 1991. The stringent response in Streptomyces coelicolor A3(2). Mol. Microbiol. 5:289-98. Tang, Y. J., H. G. Martin, P. S. Dehal, A. Deutschbauer, X. Llora, A. Meadows, A. Arkin, and J. D. Keasling. 2009. Metabolic flux analysis of Shewanella spp. reveals evolutionary robustness in central carbon metabolism. Biotechnol. Bioeng. 102:1161-9. Tautenhahn, R., C. Böttcher, and S. Neumann. 2007. Annotation of LC/ESIMS Mass Signals, p. 371-380, Bioinformatics Research and Development. Wishart, D. S., D. Tzur, C. Knox, R. Eisner, A. C. Guo, N. Young, D. Cheng, K. Jewell, D. Arndt, S. Sawhney, C. Fung, L. Nikolai, M. Lewis, M. A. Coutouly, I. Forsythe, P. Tang, S. Shrivastava, K. Jeroncic, P. Stothard, G. Amegbey, D. Block, D. D. Hau, J. Wagner, J. Miniaci, M. Clements, M. Gebremedhin, N. Guo, Y. Zhang, G. E. Duggan, G. D. Macinnis, A. M. Weljie, R. Dowlatabadi, F. Bamforth, D. Clive, R. Greiner, L. Li, T. Marrie, B. D. Sykes, H. J. Vogel, and L. Querengesser. 2007. HMDB: the Human Metabolome Database. Nucleic Acids Res. 35:D521-6. Yamaguchi, K., Y. Takahashi, T. Berberich, A. Imai, A. Miyazaki, T. Takahashi, A. Michael, and T. Kusano. 2006. The polyamine spermine protects against high salt stress in Arabidopsis thaliana. FEBS Lett. 580:6783-8.
20
487
Figures and legends
488
Figure 1: Biosynthesis pathway for ectoine. L-aspartate-β-semialdehyde (ASA) is converted into L-2,4-
489
diaminobutyrate (DABA) by the enzyme DABA aminotransferase encoded by the ectB gene. DABA is
490
converted into Nγ-acetyl-L-2,4-diaminobutyrate (ADABA) by DABA acetyltransferase encoded by the
491
ectA gene and L-ectoine (Ect) is formed from ADABA by ectoine synthetase encoded by the ectC gene (8,
492
33, 35). Ectoine can then be hydroxylated to form 5-hydroxyectoine (EctOH) by the enzyme ectoine
493
hydroxylase encoded by the ectD gene (7, 17).
494 495
Figure 2: Ectoine biosynthesis disruption mutants display a salt-sensitive phenotype, which is
496
complemented by the extracellular addition of ectoines. S. coelicolor spore stocks of strain M145,
497
ectA::Tn, ectC::Tn, ectD::Tn, ∆sigB and osaB::Tn were diluted so that a spot of 2.5 µl contained
498
approximately 105, 104, 103, 102, 101 and 100 spores (from left to right). Spots were plated on SMMS
499
medium without any addition (A, left panel), with 1 M NaCl addition (A, right panel) with 1 M NaCl and
500
20 µM ectoine addition (B, left panel) with 1 M NaCl and 20 µM hydroxyectoine addition (B, right panel)
501
or with 1 M NaCl, 20 µM ectoine and 20 µM hydroxyectoine added (C). Plates were incubated at 30°C for
502
3 days in the absence of salt and 6 days in the presence of salt.
503 504
Figure 3: Detected levels of ectoine and related compounds in continuous salt stress reflect the
505
ectoine mutants. The presence of DABA, ADABA, ectoine and hydroxyectoine was determined in wild
506
type cells and in mutants lacking ectA, ectC or ectD grown under continuous salt. DABA is only detected
507
in the etcA::Tn strain, which is expected to be deficient in the conversion of DABA into ADABA and
508
therefore accumulates the precursor. ADABA, another intermediate compound, accumulates in ectC::Tn
509
and in a lesser degree in ectD::Tn as expected. The etcD::Tn strain, which is deficient in the conversion of
510
ectoine to hydroxyectoine, accumulates large amounts of ectoine, while hydroxyectoine is accumulating in
511
the M145 as expected.
512
21
513 514
Figure 4: Overview of the putatively identified metabolites detected in the global metabolome screen.
515
A total of 229 masses could be assigned putative identities based on database matching. By far the largest
516
class of detected metabolites are lipids (107 metabolites), in particular a large number of short- and medium
517
chain fatty acids. The second largest class contains amino acids and their derivatives (66 metabolites),
518
including a large number of di- and tripeptides, many of which contained proline and glycine residues.
519 520
Figure 5: Clustering analysis uncovers the salt stress response. Clustering analysis has been used to
521
group metabolites showing similar temporal behavior into clusters. For each of the 4 clusters the mean
522
behavior is shown with error bars depicting the standard deviation. Cluster 1 contains compounds that are
523
accumulating after the salt shock, including the well known osmoprotectant proline. Cluster 4 contains
524
compounds that are decreasing after the salt shock. Interestingly, these clusters are the most homogenous of
525
all clusters as indicated by the quality score, showing that the salt-responsive metabolites are reacting as a
526
single group.
527 528
Figure 6: Comparison of time-trends of putatively identified salt stress responsive metabolites. The
529
major osmoprotectant proline, as well as several other putatively identified amino acids shows an increase
530
in response to the salt shock. The detected proline/glycine-containing di- and tripeptides also accumulate in
531
response to the salt shock (ArgGlyPro is shown as a representative example), as do a few nucleotide-related
532
compounds, including adenosine. 5’-Methylthioadenosine is one of the few compounds reproducibly
533
showing the opposite pattern, an accumulation in low-salt conditions only. Histidine and its catabolite
534
urocanate decrease in all the cultures, showing no response to the salt shock. For an overview of all of the
535
20 reproducible metabolites see Supplementary Figure E. Gray/dashed line, control non-salt shocked
536
samples; black/solid line, salt shock samples taken at 0, 4, 8, and 24 hours after salt shock from each strain
537
indicated below.
22
538
Table 1: Responses of putatively identified metabolites showing reproducible behavior in response to
539
salt stress. Response
Class
Observed mass
Upregulation in response to salt shock
AA + derivatives
Peptides
Nucleotide derivatives
Downregulation
No response
AA + derivatives Nucleotide derivatives AA + derivatives
115.0633285 131.0946287 174.1116757 165.0789786 149.0510493 204.0898776 172.0847923 328.1859033 300.1797553 301.1273854 243.1219061 136.0385108 135.0544952 267.0967539 283.0916686 117.0789786 147.0531578 297.0895601
Putative identification Name Proline (Iso-)/leucine Arginine Phenylalanine Methionine Tryptophan GlyPro ArgGlyPro LysGlyPro GluGlyPro AlaGlyPro Hypoxanthine Adenine Adenosine/deoxyguanosine Guanosine Valine Glutamate 5’-methylthioadenosine
Formula C5H9NO2 C6H13NO2 C6H14N4O2 C9H11NO2 C5H11NO2S C11H12N2O2 C7H12N2O3 C13H24N6O4 C13H24N4O4 C12H19N3O6 C10H17N3O4 C5H4N4O C5H5N5 C10H13N5O4 C10H13N5O5 C5H11NO2 C5H9NO4 C11H15N5O3S
155.0694765 138.0429274 89.04767847
Histidine Urocanate Alanine
C6H9N3O2 C6H6N2O2 C3H7NO2
540 541
23
EctB
EctD
EctC
EctA
O
O H
H
COO
NH 3 H
L-aspartateB-semialdehyde (ASA)
acetyl-CoA
H
glutamate H
NH3
2-oxoglutarate
COO
NH 3 H
L-2,4-diaminobutyrate (DABA)
H
H
H
N
CoA
H N
H
H COO
NH 3 H
Ng-acetylL-2,4-diaminobutyrate (ADABA)
H2O
CH 3
H
2-oxoglutarate + O2 Fe2+
COO
N H
ectoine (Ect)
succinate+CO2
H H N
CH 3
OH
COO
N H
hydroxyectoine (EctOH)
intensity [arbitrary units]
A:
:Tn
DABA
t ec
t ec
:T C:
n t ec
:T D:
n 5 14 M
intensity [arbitrary units]
A:
:Tn
t ec
ADABA
t ec
:T C:
n t ec
:T D:
n
5 14 M
intensity [arbitrary units]
A:
:T n
Ect
t ec
t ec
:T C:
n t ec
:T D:
n
5 14 M
intensity [arbitrary units]
A:
:T n
t ec
EctOH
t ec
:T C:
n t ec
:T D:
n
5 14 M
Contaminants Other
AA, AA derivatives, di- and tripeptides
18
AA derivatives
Nucleotide derivatives 66
38
12
AA 13
18
Non proteinogenic aminoacids
23 Di- and tripeptides
3 Vitamins + derivatives
33
Diverse
107 Proline Lipids
3
Other
9
11
Glycine + proline 10 Glycine
cluster=1,size=33,quality=0.50
-
-
cluster=3,size=45,quality=0.72 -
-
-
-
o
o
-
o
o
-
-
o
-
-
-
o
-
o
o
-
-
-
o
o
-
-
o
o
-
-
o
o
o o
-
o -
-
-
o -
-
o -
-
-
-
-
-
o
-
-
-
-
-
o -
-
o
-
o -
intensity
intensity
o
-
-
-
o
o
o
-
-
-
-
o
-
o
o
o
o o
o
o
o
o
-
o
o
o o
-
-
o
o
-
o
o o
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
o
o
-
ectA:Tn ectC:Tn ectD:Tn M145 osaB:Tn ΔsigB
-
-
-
-
ectA:Tn ectC:Tn ectD:Tn M145 osaB:Tn ΔsigB
cluster=2,size=46,quality=0.73
cluster=4,size=19,quality=0.58 -
-
-
-
-
-
intensity
-
-
-
-
-
-
o -
o
o o
o
o
o o
o
o
o
o
o
o
o
o o
o
o
-
-
-
-
-
o
o
-
-
-
-
-
-
-
-
-
-
o -
o
-
-
-
-
ectA:Tn ectC:Tn ectD:Tn M145 osaB:Tn ΔsigB
-
-
-
-
o
-
-
-
o
o o o
o
-
-
-
-
o
o o
o -
o
-
-
-
-
o
-
o
-
-
-
-
o
-
-
o
-
o
o
-
-
o
o
o
-
-
o
-
o
-
-
-
-
-
intensity
-
-
o
-
o
-
o -
ectA:Tn ectC:Tn ectD:Tn M145 osaB:Tn ΔsigB
6
NH
12
H
5
O
proline
4 3 2
4
0 M145
∆sig i B
osaB::Tn
EctA t ::Tn EctC::Tn EctD::Tn M145
O
osaB::Tn ∆sig i B
12
2.5
ArgGlyPro
OH
10
NH2
2
phenylalanine
1.5 1
fold change
fold change
arginine
2 EctA t ::Tn EctC::Tn EctD::Tn
8 6 4
0.5
2 EctA t ::Tn EctC::Tn EctD::Tn
M145
0
osaB::Tn ∆sig i B
ctA t ::Tn EctC::Tn EctD::Tn M145
osaB::Tn ∆sig i B
H 2N
NH 2
50
N
N O
OH
30
OH
adenosine
20
fold change
40
10
6 5
N O H3 C
S
HO
OH
5’-methylthioadenosine
4 3 2 1 0
EctA t ::Tn EctC::Tn EctD::Tn M145
N N
N HO
N
7
N
fold change
OH NH 2
6
0
0
O
N H
8
1
0
H2 N
10
fold change
fold change
OH N H
osaB::Tn ∆sig i B
ctA t ::Tn EctC::Tn EctD::Tn M145
osaB::Tn ∆sig i B O
O N N
OH OH
histidine
1 0.8 0.6 0.4
HN
urocanate
1 0.8 0.6 0.4 0.2
0.2 0
1.2
NH 2
fold change
fold change
1.2
HN
EctA t ::Tn EctC::Tn EctD::Tn
control
M145
osaB::Tn ∆sig i B
salt shock
0
EctA t ::Tn EctC::Tn EctD::Tn M145
osaB::Tn ∆sig i B