Applied and Environmental Microbiology-2021-Michels-AEM.00200-21.full

AEM Accepted Manuscript Posted Online 14 May 2021
Appl Environ Microbiol doi:10.1128/AEM.00200-21
Copyright © 2021 American Society for Microbiology. All Rights Reserved.

Title: Amino acid analog induces stress response in marine Synechococcus

2
3

Authors: Dana E. Michels1, Brett Lomenick2, Tsui-Fen Chou2, Michael J. Sweredoski2, Alexis

4

Pasulka1*

5
6

1

7

93407, USA

8

2

9

Engineering, California Institute of Technology, Pasadena, CA 91125, USA

Biological Sciences Department, California Polytechnic State University, San Luis Obispo, CA

Proteome Exploration Laboratory, Beckman Institute, Division of Biology and Biological

10
11

*

Corresponding author:

12
13

ABSTRACT

14

Characterizing the cell-level metabolic trade-offs that phytoplankton exhibit in response to

15

changing environmental conditions is important for predicting the impact of these changes on

16

marine food web dynamics and biogeochemical cycling. The time-selective proteome-labeling

17

approach, bioorthogonal noncanonical amino acid tagging (BONCAT), has potential to provide

18

insight into differential allocation of resources at the cellular level, especially when coupled with

19

proteomics. However, the application of this technique in marine phytoplankton remains limited.

20

We demonstrate that the marine cyanobacteria Synechococcus sp. and two groups of eukaryotic

21

algae take up the modified amino acid L-homopropargylglycine (HPG), suggesting BONCAT

22

can be used to detect translationally active phytoplankton. However, the impact of HPG

23

additions on growth dynamics varied between groups of phytoplankton. Additionally, proteomic

1

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

1


analysis of Synechococcus sp. cells grown with HPG revealed a physiological shift in nitrogen

25

metabolism, general protein stress, and energy production, indicating a potential limitation for

26

the use of BONCAT in understanding the cell-level response of Synechococcus sp. to

27

environmental change. Variability in HPG sensitivity between algal groups and the impact of

28

HPG on Synechococcus sp. physiology indicates that particular considerations should be taken

29

when applying this technique to other marine taxa or mixed marine microbial communities.

30
31

IMPORTANCE

32

Phytoplankton form the base of the marine food web and substantially impact global energy and

33

nutrient flow. Marine picocyanobacteria of the genus Synechococcus comprise a large portion of

34

phytoplankton biomass in the ocean and therefore are important model organisms. The technical

35

challenges of environmental proteomics in mixed microbial communities have limited our ability

36

to detect the cell-level adaptations of phytoplankton communities to a changing environment.

37

The proteome labeling technique, bioorthogonal noncanonical amino acid tagging (BONCAT),

38

has potential to address some of these challenges by simplifying proteomic analyses. This study

39

explores the ability of marine phytoplankton to take up the modified amino acid, L-

40

homopropargylglycine (HPG), required for BONCAT, and investigates the proteomic response

41

of Synechococcus to HPG. We demonstrate cyanobacteria can take up HPG, but also highlight

42

the physiological impact of HPG on Synechococcus, which has implications for future

43

applications of this technique in the marine environment.

44
45

2

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

24


47

INTRODUCTION
Phytoplankton are critical components of marine ecosystems, accounting for half of the

48

planet’s primary productivity and influencing global nutrient cycling (1, 2). Ecological trade-offs

49

exhibited by marine phytoplankton to maximize growth (e.g., compete for nutrients) and

50

minimize loss (e.g., increase predation defenses) shape phytoplankton diversity and community

51

structure in the ocean, which in turn significantly impacts food web dynamics and

52

biogeochemical cycling (3). A cell-level approach is necessary to elucidate how these important

53

organisms allocate resources, especially in response to changing environmental conditions.

54

The advent of ‘omics approaches has revolutionized our ability to describe marine

55

phytoplankton communities and their mechanisms of adaptation to changing environments at the

56

cellular level (4). Proteomic approaches allow for the detection of real-time changes in

57

phytoplankton physiology and metabolic state because they provide a deep insight into the

58

changes of a large fraction of the entire protein complement of the cell. Proteomics offers an

59

advantage over transcriptomics in that not all transcripts are translated into functional proteins.

60

However, due to the complexity of marine proteomes, there are challenges when applying these

61

tools to explore natural communities (5). The time-selective proteome-labeling approach,

62

bioorthogonal noncanonical amino acid tagging (BONCAT), has the potential to address some of

63

these challenges. BONCAT relies on pulse-labeling organisms with noncanonical amino acids

64

(NCAAs) like the methionine (Met) surrogate L-homopropargylglycine (HPG), such that

65

NCAAs get incorporated into nascent proteins by the cell’s endogenous translational machinery

66

(6, 7). While a number of modified amino acids successfully compete with native amino acids,

67

few are able to exploit the promiscuity of the endogenous translational machinery without

68

modification to the host cell (8). Met analogs are particularly useful because the enzyme that

3

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

46


catalyzes the esterification of Met with its tRNA, methionyl-tRNA synthetase, has low

70

specificity causing misrecognition and misincorporation of such analogs in place of methionine

71

(9). Met analogs can even serve as the initiator of translation without disruption of the

72

translational machinery (10). BONCAT has been applied successfully to visualize and identify

73

translationally active cells in natural communities (11). Furthermore, the BONCAT technique

74

offers the ability to separate the labeled proteome from the bulk proteome by exploiting the

75

chemical handle of the NCAA incorporated into the newly synthesized proteins, thereby

76

simplifying proteomic analysis and reducing challenges often encountered in natural proteomics

77

(12, 13).

78

BONCAT has been applied in a variety of cultured microorganisms and ecosystems (6, 7,

79

14), but its use in marine planktonic microbial communities has been limited. Initial studies

80

demonstrated uptake of NCAAs by marine heterotrophic bacteria and the ability to quantify

81

protein synthesis rates in natural populations (15, 16). Two additional studies showed uptake of

82

NCAAs by the eukaryotic phytoplankton Emiliania huxleyi (17) and the heterotrophic

83

flagellate Cafeteria burkhardae (18). However, none of these studies have moved beyond the

84

visualization of NCAA uptake via fluorescence microscopy to capture the protein-level

85

physiological response of these marine microbial populations. The BONCAT approach,

86

particularly when coupled with proteomics, has potential to provide insight into the differential

87

allocation of resources at the cellular level, allowing us to explore the trade-offs marine

88

microorganisms make in response to changing environmental conditions.

89

A major assumption of the BONCAT approach is that the uptake and utilization of

90

NCAAs by cells does not impact cellular physiology. Most studies report that additions of

91

NCAAs have minimal impact on cells when examined by microscopy (17, 19, 20). At the protein

4

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

69


level, no changes to protein expression or degradation were observed due to NCAA additions

93

(human embryonic kidney cells; 7 and E. coli; 14); however, one study reported alterations to

94

protein abundance (Vibrio harveyi; 9). Recent work demonstrated that NCAAs cause mild

95

perturbations to the metabolome of E. coli and that this impact was intensified when NCAAs

96

were added under stressful conditions (e.g., heat stress; 21). However, our understanding of the

97

limited impact of NCAAs on cells comes mostly from studies involving heterotrophic bacteria

98

and these assumptions may not be valid for autotrophic phytoplankton. To address this

99

knowledge gap, we explored the use of BONCAT in Synechococcus sp., a globally important

100

marine cyanobacteria. We characterized the growth of Synechococcus sp. under a range of HPG

101

concentrations and optimized the fluorescence signal to detect this uptake via epifluorescence

102

microscopy. In addition, we examined changes in protein expression of Synechococcus sp.

103

grown with HPG under normal and nitrate-stressed conditions relative to a non-HPG control.

104

Finally, we characterized the growth and quantified HPG uptake under a range of HPG

105

concentrations in two eukaryotic phytoplankton models, Ostreococcus sp. and Micromonas

106

pusilla, to test whether they exhibited the same initial sensitivity to HPG additions as

107

Synechococcus sp.

108
109
110

RESULTS

111

Impact of HPG concentration on phytoplankton growth dynamics

112

Phytoplankton exhibited different sensitivities to HPG concentration. For both eukaryotic

113

green algal models (M. pusilla and Ostreococcus sp.), growth in the presence of HPG was

114

similar to that exhibited in the negative (e.g., non-HPG) control for all HPG concentrations

5

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

92


tested (up to 100 µM; Figure 1A, B). However, for Synechococcus sp., HPG concentrations as

116

low as 25 µM disrupted normal growth dynamics (Figure 1C). Compared to the maximum

117

growth exhibited by the negative control, 25 µM HPG additions reduced Synechococcus sp.

118

growth by 39%, whereas 50 µM HPG additions reduced growth by 51%. 100 µM HPG

119

concentrations (the concentration used in previous studies with marine heterotrophic bacteria in

120

culture; 13) resulted in a complete crash of the Synechococcus sp. culture (data not shown).

121

Synechococcus sp. grew normally and reached the same maximum growth when exposed to 10

122

µM HPG and lower concentrations. Synechococcus sp. growth with 10 µM HPG additions was

123

characterized three additional times to confirm this result (data not shown).

124
125

Fluorescence detection of BONCAT signal

126

Epifluorescence microscopy was used to detect and visualize HPG incorporation by the

127

phytoplankton from HPG-growth experiments using the highest HPG concentration that did not

128

alter cell growth dynamics (Figure 1; 100 µM for M. pusilla and Ostreococcus sp. and 10 µM for

129

Synechococcus sp.). As outlined in the methods section Microscopy and Image Analysis,

130

heterotrophic bacteria present in the cultures were excluded from the data prior to interpretation

131

of HPG incorporation by the phytoplankton cells (Figures S1 and S2). While the proportion of

132

bacteria in the cultures could not be determined for Ostreococcus sp. and M. pusilla (because

133

bacteria were visually removed during ROI selection), the proportion of bacteria present in

134

Synechococcus sp. cultures over multiple experiments ranged from 17-23% of the population.

135

For all taxa, the fluorescence signal in the blue (e.g., DAPI-stained cells) and red or

136

orange (e.g., autofluorescence from phytoplankton pigments) channels were consistent between

137

the negative and positive treatments, providing visual evidence that the cells appeared healthy

6

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

115


and intact when grown in the presence of HPG (Figures 2,3,4). A bright signal in the green

139

channel (e.g., fluorescence signal of the azide-containing CR-110 fluorophore) was visually

140

apparent in cultures amended with HPG relative to negative control cultures (Figures 2,3,4).

141

Quantitative analysis based on the green fluorescence intensity revealed that the HPG-amended

142

treatments (positive) were significantly different from the negative controls for all phytoplankton

143

models across all time-points (Mann-Whitney U Test; Table S1, Figure 5). However, it is

144

important to note that while Synechococcus sp. cells grown with HPG were still brighter in

145

comparison to HPG negative cells, these cells exhibited some autofluorescence in the green

146

channel due to the presence phycobiliproteins hence the greater overlap in the signal between the

147

positive and negative treatments (Figure 2, 5C). M. pusilla exhibited the strongest fluorescence

148

signal as a result of HPG incorporation at 48 h (Figure 5A). In contrast, the fluorescence signal

149

as a result of HPG incorporation (e.g., CR-110) in Synechococcus sp. and Ostreococcus sp.

150

increased over time (Figure 5B, C) and exhibited the strongest fluorescence relative to the

151

negative control at 72 h post HPG addition.

152
153

Physiological response of Synechococcus sp. to HPG additions under replete and nutrient-

154

limited conditions

155

Following resuspension of cultures in the appropriate treatment media with HPG (Table

156

2), replete cultures exhibited a typical growth rate, while nitrate-limited cultures exhibited a

157

reduced growth rate (Figure 6). Control cultures (e.g., no spin conditions, no HPG) continued to

158

increase exponentially. Epifluorescence microscopy revealed consistent fluorescence signals in

159

the blue (e.g., DAPI-stained cells) and orange (e.g., autofluorescence from phycobiliproteins

160

pigments) channels across treatments and time points. However, green fluorescence intensity

7

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

138


(fluorescence signal of the azide-containing CR-110 fluorophore) was visually brighter for both

162

replete and nitrate-limited HPG-amended treatments compared to the control (Figure 7). This

163

visual difference in green fluorescence intensity was quantitatively and significantly different in

164

HPG-amended treatments (replete and nitrate-limited) relative to the control across both time-

165

points (Table S2, Figure S1).

166

Proteomic analysis collected 22,209 MS2 spectra, from which 5,969 peptide-to-spectrum

167

matches and 5,551 peptide groups were identified. A total of 1,033 proteins were identified,

168

(representing about 36% of all predicted proteins in Synechococcus sp. strain CC9311; 19) and

169

496 quantified proteins were used for analysis after filtering (see Methods section ‘Proteomics’).

170

HPG labeling was detected in 68 of the final quantified proteins. Relative to the negative control,

171

222 proteins had significantly different expression (i.e., a minimum log2 fold-change of 1 and an

172

adjusted p-value less than 0.05) in the HPG positive treatments relative to the negative control

173

(Figure 8, Table S3). Of these proteins, 122 were up-regulated and 100 were down-regulated.

174

HPG labeling was detected in 25 of these significant proteins (Table S5). The proteins

175

differentially expressed in the nitrate-limited condition vs. the control had a greater median log2

176

fold-change value than the proteins differentially expressed in the nitrate-replete condition vs.

177

the control (Figure S2), suggesting that the added stress of nitrate limitation magnified the

178

proteomic changes caused by HPG treatment.

179

Proteomic analysis revealed that HPG influenced several aspects of metabolism including

180

nitrogen metabolism, general protein stress, and energy production (Table 3). Glutamate

181

synthetase (ferredoxin-dependent glutamate synthase, Fd-GOGAT) and glutamine synthetase

182

(glnN) were significantly upregulated in HPG-treatments compared to the control. Chaperonin

183

proteins (groEL1, groEL2, dnaK and htpG) and a probable cytosol aminopeptidase (pepA) were

8

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

161


also significantly upregulated in HPG treatments. Many antenna proteins were upregulated in

185

HPG treatments, including phycobilisome linker polypeptides (cpcC, cpcD, cpeC, cpeD1,

186

cpeD2), phycobilisome rod-core linker polypeptides (cpcG1, cpcG, apcE), allophycocyanin

187

apoproteins (apcB1, apcD), and phycoerythrin chain peptides (cpaA2, cpaB1, cpaB2). Two other

188

antenna proteins (cpeT, cpeS) were significantly downregulated in HPG treatments relative to

189

the control. Major components of photosystem I were upregulated in HPG treatments relative to

190

the control, including psaA, psaB, psaC, psaD, psaF, and psaL. Some proteins of photosystem II

191

were also upregulated (psbC, psbJ, psbW and psbV). Two major proteins of ATP synthase were

192

upregulated (atpH and atpF), while one protein was downregulated (atpD). Many peptides

193

related to the phycobilisome and photosystem I were labeled with HPG including, ferredoxin,

194

psaC, psaF, cpcD, cpa2, cpaB1, and cpaB2. Additionally, one of the ATP synthase F1

195

subcomplex beta subunits (atpD) was HPG labeled. The antioxidant proteins peroxiredoxin and

196

thioredoxin-dependent thiol peroxidase, prx and prxQ, respectively were also upregulated in

197

HPG treatments.

198

When comparing protein expression between the two HPG positive treatments (nitrate-

199

replete vs. nitrate-limited), 14 proteins were significantly different between these two groups. Of

200

these proteins, 4 were up-regulated and 10 were down-regulated (Table S5). HPG labeling was

201

detected in 4 of these significant proteins. However, for 13 of these 14 proteins, the directionality

202

of the log2 fold-changes for this pairwise comparison was the same as the directionality in the

203

pairwise comparison of HPG positive treatments relative to the control. Therefore, the

204

identification of these proteins as being differentially expressed between nitrate-replete vs.

205

nitrate-limited treatments largely reflects the greater magnitude of log2 fold-change values in the

206

nitrate-limited treatment rather than a meaningful biological difference in the proteome of the

9

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

184


two conditions (Table S5, Figure S2). These proteins predominantly indicate that when a cell is

208

under nutrient stress, the addition of HPG exhibits a greater interference on nitrogen metabolism

209

(upregulation of the nitrogen-responsive regulatory protein [ntcA] and glutamine synthetase [GS]

210

and downregulation of ferredoxin-nitrite reductase [Fd-nir]) and energy production (upregulation

211

of NADH dehydrogenase and downregulation of light-independent protochlorophyllide

212

reductase iron-sulfur ATP-binding protein [chlL]; TableS5).

213
214
215

DISCUSSION

216

The BONCAT technique has helped scientists begin to address key questions in

217

microbial ecology. This approach provides a tool to link microbial function with phylogeny by

218

identifying translationally active cells in cultures and natural communities (14, 16). BONCAT

219

has recently been applied to study marine bacterioplankton and obtain single-cell protein

220

synthesis rates (15), but its use in marine phytoplankton communities remains limited. When

221

coupled with proteomics, this technique has the potential to help elucidate the mechanisms by

222

which marine microorganisms adapt and survive in changing environmental conditions (e.g.,

223

Pseudomonas aeruginosa, 13; Vibrio harveyi, 23). In this study we demonstrate that the marine

224

cyanobacteria Synechococcus sp. and two groups of eukaryotic algae can take up the modified

225

amino acid, HPG. Overall, our findings suggest that BONCAT can be used to detect

226

translationally active phytoplankton. However, among different phytoplankton groups, we

227

observed variability in how HPG impacted normal growth dynamics (Figure 1). Furthermore,

228

despite normal growth patterns when exposed to 10 µM HPG concentrations, variations in

229

protein expression between Synechococcus sp. in HPG treated cultures vs. non-HPG control

10

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

207


cultures revealed an influence of HPG on cyanobacterial cell physiology (Figure 8). Therefore,

231

the ability to use BONCAT as a tool for understanding the cell-level response to stressors may be

232

limited in Synechococcus sp. and other cyanobacteria.

233

While there is evidence to suggest that some phytoplankton take up amino acids, the field

234

lacks consensus on the prevalence, rates, and occurrences of amino acid uptake as well as the

235

importance of these amino acids as a nitrogen source for phytoplankton (24, 25). Visualization of

236

fluorescently-labeled cells after exposure to HPG suggest that the phytoplankton tested in this

237

study have the ability to take up free amino acids (Figures 2, 3, 4, 5). This finding is consistent

238

with previous work demonstrating that the coccocolithophorid Emiliania huxleyi and marine

239

flagellate Cafeteria burkhardae can take up HPG (17, 18), and that Synechococcus sp. can take

240

up organic compounds including amino acids (27, 28). Laboratory studies with cultured

241

organisms have demonstrated that some phytoplankton can grow successfully on certain amino

242

acids as a sole nitrogen source (28, 29), but conclusions from field studies with natural

243

populations have been more variable (25, 29). However, studies on marine picocyanobacteria

244

have shown natural population take up various amino acids (30-34). While most studies assume

245

amino acids are taken up intact, some phytoplankton possess enzymes that oxidize L-amino acids

246

at the cell surface to obtain ammonium for cellular uptake (29, 34). These cell surface

247

deaminases are responsible for decomposing amino acids, effectively separating the nitrogen

248

source from the carbon backbone and allowing for uptake (36). Therefore, further work is needed

249

to examine amino acid uptake mechanisms for different groups of phytoplankton such that we

250

can determine if this technique can be more broadly applied to natural phytoplankton

251

communities.

11

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

230


Overall, these culture-based experiments revealed that phytoplankton exhibit varied

253

sensitivities to HPG concentration (Figure 1). The micro-eukaryotes Ostreococcus sp. and M.

254

pusilla appeared to be less sensitive to HPG than Synechococcus sp., such that their growth (as

255

measured via spectrophotometry) was not inhibited by HPG concentrations as high as 100 µM

256

(Figure 1A, B). The decrease in growth observed in Synechococcus sp. at 25 µM HPG

257

concentrations suggests that this group experienced physiological perturbations with the addition

258

of this modified amino acid (Figure 1C). Very few studies have actually investigated the impact

259

of HPG on an organism’s cellular physiology. Recent work by Steward et al. (21) revealed that

260

HPG additions caused a shift in the metabolome of E. coli that was intensified by heat stress.

261

However, different ecosystems and organisms are likely to have varied sensitivities to HPG

262

additions; therefore, knowledge from one model organisms (e.g., E. coli), may not be applicable

263

to other model organisms or ecosystems. In this study, we found that Synechococcus sp. cells

264

exposed to 10 µM HPG exhibited protein-level changes to nitrogen metabolism, general protein

265

stress, and energy production (Figure 8, Table 3), even though growth was not inhibited until the

266

cells were exposed to 25 µM of HPG. In agreement with Steward et al. (21), the changes in

267

protein expression caused by HPG were intensified (greater log2 fold-change) when the cells

268

were under nutrient stress (Figure S2). Interestingly, many of the proteins that had significantly

269

altered expression were also labeled with HPG (Table 3). Furthermore, HPG-labeled proteins

270

predominantly occurred within structural complexes (e.g., ATP synthase, phycobilisome).

271

However, it is important to note that due to the strong correlation between protein abundance and

272

sequence coverage obtained in bottom-up proteomics (37), we are more likely to see HPG

273

incorporation in higher abundance proteins (e.g., structural and metabolic housekeeping genes).

274

At this time, it is unclear whether the increased expression of proteins in HPG-treated cells

12

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

252


would lead to increased products and affect metabolic pathways, or whether the HPG-

276

substitutions caused structural or functional issues with those proteins. In the latter case, the

277

increased expression could have resulted from the cells replacing polypeptides in malfunctioning

278

HPG-labeled proteins, particularly for key structural complexes. Below, we explore the

279

metabolic consequences of HPG addition to Synechococcus sp. cells based on significant

280

changes we detected at the protein level.

281

The upregulation of glutamine synthetase (GS) and glutamate synthetase (GOGAT) in

282

HPG-treated cultures (Table 3) indicates potential inhibition of these enzymes by HPG and

283

therefore key nitrogen cycling processes in the cell. In cyanobacteria, ammonium is incorporated

284

into carbon skeletons by the GS/GOGAT cycle (38). Glutamate and glutamine produced in this

285

cycle play important roles in the distribution of nitrogen throughout the cell to other nitrogenous

286

compounds. Glutamate is not only the direct precursor for some amino acids and 5-

287

aminolevulinate (the immediate precursor for phycobilin, chlorophyll and porphyrin

288

biosynthesis), but it also functions as the primary nitrogen donor for the synthesis of other

289

nitrogen-containing metabolites (38). Specific inhibition of glutamine synthetase is well-

290

documented in another strain of Synechococcus sp. by doping with the NCAA, L-methionine-

291

sulfoximine (MSX; 39, 40). When ammonium was provided as the sole nitrogen source, MSX

292

acted as a specific inhibitor of GS thereby inhibiting ammonium assimilation, essentially

293

mimicking nitrogen starvation (40). This evidence of MSX’s functioning in other Synechococcus

294

sp. strains supports the idea that HPG may be inhibiting these critical enzymes of nitrogen

295

cycling in Synechococcus sp.

296
297

A general stress response, with particular evidence for protein stress, in the presence of
HPG was indicated by the upregulation of chaperonins. The upregulation of chaperonins is well-

13

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

275


documented as part of the core general stress response in cyanobacteria (41). Chaperonin

299

proteins are thought to transiently bind polypeptides that have been forced to take non-native

300

conformations under stressful conditions. Through this binding, chaperonins function to

301

temporarily hold the non-native polypeptides and prevent aggregation in the crowded

302

environment of the cell (41). Chaperonins with increased expression in HPG dosed

303

Synechococcus sp. cells included groEL (1 and 2), dnaK, and htpG (eukaryotic homologs are

304

heat shock proteins hsp60, hsp70 and hsp90, respectively; 38; Table 3). DnaK also functions to

305

solubilize aggregates of partially denatured proteins and rescue the polypeptides (41).

306

Additionally, upregulation of the cytosol aminopeptidase (pepA), which is involved in protein

307

regulation and turnover indicates cellular stress, as this protein is upregulated in E.coli under heat

308

stress (42). Together, the significant upregulation of this suite of proteins provides evidence that

309

HPG additions lead to protein misfolding and stress in Synechococcus sp. cells.

310

Collectively the upregulation of proteins associated with the light harvesting capacity of

311

Synechococcus sp. (phycobilisome antenna proteins and photosystem I) and oxidative

312

phosphorylation indicate that cells are experiencing an interference with energy production under

313

HPG addition (Table 3). The phycobilisome is the major light harvesting apparatus of

314

cyanobacteria. This highly ordered supramolecular complex is made up of linker polypeptides

315

and phycobiliproteins with unique spectral properties (phycoerythrin [PE, Amax = 560 nm],

316

phycocyanin [PC, Amax = 620 nm] and allophycocyanin [AP, Amax = 650 nm]; 33). Linker

317

polypeptides stabilize the phycobilisome structure and ensure optimal functioning by

318

encouraging unidirectional flow of energy from the periphery to the core, and later to the

319

photosynthetic reaction center (43, 44). The upregulation of many proteins related to the

320

phycobilisome, especially linker polypeptides, indicates potential instability or disfunction of the

14

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

298


phycobilisome in HPG treated cells. This upregulation could also indicate attempts to increase

322

light harvesting capacity for energy production in cells treated with HPG. Additionally, some of

323

the antenna proteins related to phycoerythrin (cpa2, cpaB1 and cpaB2) that were significantly

324

upregulated in HPG treatments were also labeled with the NCAA. The hptG protein was also

325

significantly upregulated in HPG dosed treatments, providing evidence of non-native protein

326

folding or aggregation of linker polypeptides in HPG treated cells. Interestingly, the stress

327

response chaperonin, htpG, has been shown to interact with and stabilize a 30 kDa rod linker

328

polypeptide of the phycobilisome in a different strain of Synechococcus (45). Together, the

329

observed changes to phycobilisome-related proteins indicates potential issues in stability or

330

function of the major light harvesting complex in HPG treated Synechococcus sp.

331

Photosystem I (PS I) in cyanobacteria is a light driven reaction center that changes the

332

energy of a photon into free chemical energy though oxidation of cytochrome c6 or plastocyanin

333

and reduction of ferredoxin or flavodoxin (46). Major components of PS I that were upregulated

334

in HPG treatments relative to the control included psaA and psaB (heterodimer of the integral

335

reaction center), psaC (which provides a path for the electrons out of the membrane phase and to

336

the stromal phase, allowing ferredoxin to be reduced with high quantum efficiency), psaD

337

(which docks ferrodoxin or flavodoxin), psaF (which docks plastocyanin or cytochrome c6), and

338

psaL (the connecting protein for PS I trimers and state transitions). While the cause of

339

Synechococcus sp. upregulating these critical PS I proteins in the presence of HPG remains

340

unknown, we hypothesize this is due either to an attempt to increase the energy production by

341

this reaction center or to cope with structural or functional issues caused by HPG substitution.

342

A series of electron transfer reactions harvest energy in cyanobacteria to create a

343

transmembrane electrochemical proton gradient that is used to drive the synthesis of ATP by an

15

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

321


F-type ATPase, allowing energy to be temporarily stored and easily accessed by many enzymes

345

through the cell (47). ATP synthase genes examined in select cyanobacteria are arranged in two

346

gene clusters, atp1 (codes genes atpI-atpH-atpG-atpF-atpD-atpA-atpC) and atp2 (codes

347

remaining genes atpB-atpE; 48). Two major proteins of cluster 1 were upregulated (atpH and

348

atpF), while the following protein was downregulated and also labeled with HPG (atpD),

349

indicating that in the presence of HPG, Synechococcus sp. are likely experiencing interference

350

with this important energy production complex.

351

The disruption to normal energy production systems described above suggests that HPG

352

could indirectly cause oxidative stress in Synechococcus sp. Oxygen is a powerful electron

353

acceptor yet its intermediates, which are generated through photosynthesis and electron

354

transport, can have highly damaging effects on metabolic networks (49). Imbalances in the

355

generation of reactive oxygen species and antioxidant responses, lead to oxidative stress, which

356

commonly occurs due to environmental stressors (e.g., UV stress, nutrient stress). The

357

upregulation of two antioxidant proteins, peroxiredoxin (prx) and thioredoxin-dependent thiol

358

peroxidase (prxQ), which are important proteins for maintaining redox homeostasis in

359

Synechococcus sp., provide evidence of redox imbalance in HPG treated cells (50, 51).

360

Overexpressed prxQs have shown protection from oxidative stress in cyanobacteria (51). Overall

361

changes to light harvesting and energy production pathways in concert with upregulation of

362

antioxidant proteins indicates that HPG may indirectly cause oxidative stress in Synechococcus

363

sp.

364

Phytoplankton experience a consistently changing set of environmental conditions in the

365

ocean (52). Therefore, investigating the impact of NCAAs on microbial populations in

366

conjunction with an environmental stressor provides a more realistic evaluation of the technique

16

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

344


(21). Populations of marine cyanobacteria are especially influenced by nutrient availability (53,

368

54) and as such, nitrate limitation was a physiologically appropriate stressor to test in

369

combination with NCAA additions. Further, there is supporting literature at the transcript level

370

describing the response of Synechococcus sp. to nitrate stress (55, 56). These transcript level

371

responses typically include the reduction of photopigments and photosynthetic capacity as well

372

as the upregulation of the nitrogen control regulator genes (56). Ideally, this investigation would

373

have led to a better protein-level understanding of nitrate stress in Synechococcus sp. However,

374

the inference with nitrogen cycling (as well as other core stress responses) induced by the

375

addition of HPG, limits our ability to determine if the protein expression in nitrate-limited cells

376

can be attributed to nitrogen stress or if the nitrogen stressed cells were more susceptible to the

377

issues induced by HPG additions. Therefore, follow up experiments without HPG could provide

378

insight into the protein-level responses of Synechococcus sp. to environmental stressors and

379

allow for better comparison of the protein and transcript-level responses.

380

Metabolic labeling of proteins with NCAAs provides a unique tool for characterizing

381

translationally active microbial communities (14). However, if HPG directly impacts the

382

physiology of cells, as we demonstrate here for Synechococcus sp., the use of this technique to

383

elucidate the protein-level physiological response of microorganisms to changing environmental

384

conditions may be limited. While Synechococcus sp. treated with 10µM of HPG exhibited the

385

same growth dynamics as non-HPG controls (Figure 1C), we observed significant and

386

potentially detrimental changes in important pathways (Table 3). At this time, it is unclear if and

387

how these protein-level changes translated into altered metabolic rates, since growth under this

388

concentration of HPG appeared normal. However, these effects were likely intensified at higher

389

concentrations of HPG, such that when exposed to 25 µM of HPG, Synechococcus sp. did

17

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

367


exhibit a reduction in growth. These findings highlight the importance of coupling HPG

391

additions with several biologically relevant rate measurements (e.g., growth, photosynthetic

392

efficiency) to determine if and to what extent HPG alters metabolic rates. When applied in a

393

natural setting, these types of HPG-induced effects could alter interactions between organisms.

394

For example, in the marine environment, there is close coupling between phytoplankton and

395

heterotrophic bacterial growth dynamics (57, 58), because heterotrophic bacterial growth in the

396

euphotic zone is strongly influenced by phytoplankton-derived dissolved organic matter (59-61).

397

Marine Synechococcus sp. and heterotrophic bacteria are known to be spatially and ecologically

398

intimate in natural settings and in culture, frequently conjoining or forming networks (62, 63).

399

Given the tight interaction of these organisms, both physically and for essential carbon and

400

nutrients, changes in one organism’s physiology, (e.g., Synechococcus sp.) in response to HPG,

401

could have cascading effects on the heterotrophic bacteria that exist in community with those

402

cells.

403

The BONCAT technique lends itself to investigating a wide range of ecological

404

questions. Therefore, if the goal of the study is to visualize and sort translationally active marine

405

microorganisms (64) this technique may be appropriate. In contrast, if the goal of the application

406

is to characterize and track the physiology and proteomic response of marine microorganisms

407

including cyanobacteria to changing environmental conditions, this approach may not be

408

appropriate. It is however, important to recognize that the HPG concentrations used in this study

409

were meant to maximize fluorescence intensity in order to detect HPG uptake in cyanobacteria.

410

These higher concentrations may not always be necessary for other ecosystems and ecological

411

questions. Additionally, it would be worth exploring other commonly used NCAAs, such as L-

412

azidohomoalanine (AHA), in these types of sensitivity experiments. AHA has been applied to an

18

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

390


equally wide range of study systems (7, 20, 65-67) and demonstrates potentially less toxicity and

414

disturbance to the physiology of the study system (21), although its impact has not been

415

investigated at the proteomic level. Overall, our results highlight an imperative aspect of

416

investigation for studies utilizing HPG to ensure the NCAA addition does not influence cell

417

metabolism. Different taxa are likely to show varying sensitivities to NCAA additions and these

418

considerations must be taken into account when applying this approach to study marine

419

microbial communities.

420
421
422

METHODS

423

Culture conditions

424

Experiments were conducted with three different algal cultures including one

425

cyanobacteria and two green algae – Synechococcus sp. (strain CC9311; CCMP3074; 68),

426

Micromonas pusilla (CCMP487) and Ostreococcus sp. (MBIC10636), respectively. Cultures

427

were maintained in L1 medium minus silica (69) using salt solutions from ESAW Medium rather

428

than filtered natural seawater (70, 71). Cultures were maintained in an incubator at 18C on a

429

12:12 h light:dark cycle, however, Ostreococcus sp. and M. pusilla were grown under higher

430

light (75 µmoles photons m-2 s-1) relative to Synechococcus sp. (20-25 µmoles photons m-2 s-1).

431

All cultures were transferred and maintained under sterile conditions, but were not axenic.

432

Heterotrophic bacteria were present, but in low densities relative to the algal strains (see Results

433

section Microscopy and Image Analysis for details).

434
435

HPG sensitivity experiments

19

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

413


The three organisms were grown to exponential phase according to the conditions

437

described above. L-homopropargylglycine (HPG; Click Chemistry Tools), a methionine

438

analogue, was resuspended in 0.2 µM sterile filtered water and then added to the cultures in

439

exponential phase at a range of concentrations between 0.2 to 100 µM (Table 1). For M. pusilla

440

and Ostreococcus sp., all HPG concentrations were tested in a single experiment. However, for

441

Synechococcus sp., HPG concentrations were tested over the course of three experiments (high

442

[100µM], mid-range [10, 25 and 50µM] and low [0.2, 0.5, 1 and 5µM]) to capture the maximum

443

concentration of HPG that could be added without influencing cell growth dynamics. Across

444

these experiments we added HPG during exponential growth phase (OD450 range between 0.2 –

445

0.3).

446

Cultures were monitored daily by optical density (OD450; 72, 73). Standard curves

447

comparing spectrophotometric absorbance at 450 nm and cell counts (obtained by

448

epifluorescence microscopy) showed strong correlations for all cultures (r2 values;

449

Synechococcus sp. = 0.81, M. pusilla = 0.99, Ostreococcus sp. = 0.91). Samples for microscopy

450

were taken at 24, 48 and 72 h after the addition of HPG and fixed with 2% paraformaldehyde

451

(PFA) at 4C. Samples were in fixative for 24 to 48 h, centrifuged, washed in 1X PBS,

452

resuspended in 50:50 v/v EtOH:H2O, and stored at -20C. Line plots were generated in R (74)

453

using the package ggplot2 (75).

454
455
456

Impact of HPG addition with and without nutrients
To explore the impact of HPG on cell growth and physiology with and without nitrate,

457

cultures of Synechococcus sp. were amended with HPG in combination with nutrient stress (via

458

nitrate removal from culture medium). Synechococcus sp. cultures were maintained in

20

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

436


exponential phase in sterile L1 medium minus silica. When culture density reached OD450 ~0.4

460

(~1.5 x 108 cells/ml), all experimental cultures were centrifuged and cells were washed in L1

461

medium without nitrogen (following nitrogen limitation protocol of 56). Concentrated cultures

462

were resuspended to half of the starting OD450 in fresh L1 medium, either with nitrate (replete) or

463

without nitrate (nitrate-limited) (Table 2). These experimental cultures were amended with 10

464

µM HPG at the time of set up, as this concentration was determined not to affect cell growth by

465

initial HPG sensitivity experiments (figure 1C). Control cultures were maintained with replete

466

nutrients and no HPG. In order to not disrupt the exponentially growing population within the

467

control, the control was not centrifuged. A separate centrifuge vs. no-centrifuge control

468

experiment revealed that centrifugation had a minimal impact on the protein expression of

469

Synechococcus sp. cultures; furthermore, this comparison revealed that centrifugation did not

470

induce any changes to protein expression that we attribute to HPG in this study (Table S6). All

471

treatments were run in triplicate with a total volume of all 100 ml in 250 ml Erlenmeyer flasks.

472

All experiments were set up immediately before the onset of the light cycle and OD450 was

473

monitored daily as a proxy for cell density.

474

Preliminary analysis from the HPG-sensitivity experiments revealed that HPG uptake by

475

Synechococcus sp. was visible by fluorescence after 48 h and increased after 72 h. Therefore,

476

samples for microscopy and proteomics were taken 48 and 72 h following HPG additions. The

477

BONCAT signal showed greater intensity at 72 h compared to 48 h, indicating increased labeling

478

at the later time point, so 72 h samples were analyzed for proteomics. For microscopy, triplicates

479

of 1.8 ml from each culture were fixed with paraformaldehyde (PFA; 2% final concentration) for

480

24 h at 4C. Samples were then centrifuged, washed in 1X PBS, resuspended in 50:50 ethanol:

481

1X PBS and stored at -20C. For protein samples, 20 ml of culture were centrifuged,

21

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

459


resuspended in fresh L1 medium, either replete or lacking nitrate according to the treatment and

483

transferred to a 1.5 ml Eppendorf tube. This sample was centrifuged again and supernatant was

484

removed before fast-freezing in liquid nitrogen and storing at -80C.

485
486
487

BONCAT Click Chemistry
The incorporation of HPG into Synechococcus sp. proteins was detected via

488

epifluorescence microscopy. The click reaction, or copper(I)-catalyzed cycloaddition,

489

fluorescently labels the methionine analog by incubating it with an azide-linked fluorophore and

490

copper(II) under reducing conditions to allow copper(I) to catalyze binding of the azide to the

491

alkyne (6). Prior to conducting the click reaction, all samples were spotted on Teflon printed

492

slides (Electron Microscopy Sciences, PTFE Printed Slides) in volumes ranging from 2 – 10 µl

493

(depending on cell concentration) and air dried. Samples on slides were put though an ethanol

494

dehydration series (50:50, 80:20 then 96:4 v/v EtOH:H2O) prior to incubation with freshly

495

prepared click solution (4 µl of dye-premix [CuSO4, 0.1 mM; THPTA, 0.5 mM; CR-110-azide

496

fluorophore, 2 mM] added to 254 µl buffer solution [sodium ascorbate, 5 mM; aminoguanidine

497

hydrochloride, 5 mM; 1X PBS]). Slides were incubated for 30 minutes in the dark at room

498

temperature in a humid chamber. Post incubation samples were washed in 1X PBS and then

499

H2O. After samples were completely dry, coverslips were mounted over wells with DAPI

500

VectaShield mounting medium (Vector Labs).

501
502
503
504

Microscopy and Image Analysis
Algal samples were analyzed with a Zeiss Axio Observer Z1 Inverted Epifluorescence
Microscope using a 100X objective. Digital images were acquired with a 6-megapixel CCD

22

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

482


camera (Zeiss Axiocam 506 mono). Peak channel excitation and emissions wavelength/bandpass

506

in nm were 365 and 445/50 for blue (DAPI-stained cells), 470/70 and 525/50 for green

507

(fluorescence signal of the azide-containing CR-110 fluorophore), 550/25 and 605/70 for orange

508

(autofluorescence from phycobiliproteins), and 440/40 and 675/50 for red (chlorophyll

509

autofluorescence).

510

Images were analyzed using an in-house image analysis pipeline in MATLAB (76). In

511

short, regions of interest (ROIs) were selected by applying a signal threshold to the images. The

512

mean blue (e.g., DAPI-stained cells), green (e.g., fluorescence signal of the azide-containing CR-

513

110 fluorophore), and red (or orange) (e.g., autofluorescence from chlorophyll photopigments in

514

Micromonas pusilla and Ostreococcus sp. and autofluorescence from phycobiliprotein

515

photopigments in Synechococcus sp.) fluorescence values were recorded for each ROI and the

516

data were normalized by image exposure time. For all algal strains, heterotrophic bacteria were

517

removed during the image analysis process prior to interpretation of HPG uptake by the

518

phytoplankton cells. For Ostreococcus sp. and M. pusilla, heterotrophic bacteria were excluded

519

during ROI selection because their smaller size and absence of chlorophyll-a made them visually

520

distinct (Figure S3). For Synechococcus sp., because heterotrophic bacteria spanned the same

521

size range and were less visually distinct, they were excluded after ROI selection by applying a

522

cutoff based on the orange (e.g., representing phycoerythrin autofluorescence in Synechococcus

523

sp.) to blue (e.g., DAPI-stained cells) signal ratio and/or orange signal value alone (Figure S4).

524

After removing heterotrophic bacteria from the data, box plots of the mean green (e.g.,

525

representing the fluorescence signal of the azide-containing CR-110 fluorophore) fluorescence

526

intensity were used to visualize differences between positive and negative HPG treatments. The

527

number of ROIs included in the final analysis varied across the algal strains based on their

23

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

505


densities in each image: for Synechococcus sp., 500 - 1000 ROIs were selected per image (~5

529

images per treatment), Ostreococcus sp., 100-500 ROIs were selected per image (~3 images per

530

treatment), and for M. pusilla, 100 ROIs were selected per image (~4 images per treatment).

531

Differences in fluorescence intensity between positive and negative HPG were tested using a

532

Mann-Whitney U Test (77) Box plots were generated with custom scripts in R (74) using the

533

package ggplot2 (75).

534
535
536

Proteomics

537

Samples from Synechococcus sp. at 72 h time point were selected for proteomic analysis

538

based on greatest BONCAT signal intensity observed via microscopy. Cell pellets were lysed on

539

ice with 100 L 0.05%SDS/0.5M TEAB lysis buffer by vortexing, syringe titration with a 23G

540

needle (30x), and dounce homogenization (30x). Lysates were then clarified at 16,000 g for 5

541

min at 4C and protein concentration measured by Bradford assay. Ten microgram lysates from

542

each sample in 52 L lysis buffer were reduced with 3 mM TCEP for 1hr at 50C, alkylated with

543

10 mM iodoacetamide for 15 min at room temperature, digested with 1:100 LysC for 2 h at room

544

temperature, and digested with 1:25 trypsin overnight at 37C.

545

Digests were stopped by acidifying with 3.25 L 100% formic acid and desalted on

546

StageTips packed in-house with Empore C18 extraction material (3M cat. # 2215). Desalted

547

peptides were lyophilized, resuspended in 10 L 100 mM TEAB, and peptide quantification

548

performed with the Pierce colorimetric peptide quant assay (Thermo cat. #23275). 5 ug peptides

549

per sample were brought to 10 L total in 100mM TEAB, then labelled with 100 g TMT-11

550

(Thermo) reagents in 4 L anhydrous acetonitrile for 2 h at room temperature. TMT reactions

24

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

528


were quenched with 1 L 5% hydroxylamine for 15 minutes at room temperature, combined,

552

lyophilized, and desalted on a C4 macrotrap cartridge (Optimize Technologies #10-04818-TM)

553

on an Agilent 1100 HPLC. SDS was then removed with HiPPR detergent removal resin (Thermo

554

#88305), and the peptides were resuspended in solvent A (2% ACN, 0.2% formic acid).

555

Liquid chromatography-mass spectrometry (LC-MS) analysis was carried out on an

556

EASY-nLC1000 (Thermo Fisher Scientific, San Jose, CA) coupled to an Orbitrap Fusion Tribrid

557

mass spectrometer (Thermo Fisher Scientific, San Jose, CA). Approximately 250 ng peptides

558

were loaded onto an Aurora 25cm x 75µm ID, 1.6µm C18 reversed phase column (IonOpticks,

559

Parkville, Victoria, Australia) and separated over 136 min at a flow rate of 350 nL/min with the

560

following gradient: 2–6% Solvent B (7.5 min), 6-25% B (90 min), 25-40% B (30 min), 40-100%

561

B (1 min), and 100% B (15 min). Solvent B consisted of 80% ACN, 0.2% formic acid. MS1

562

spectra were acquired in the Orbitrap at 120K resolution with a scan range from 400-1500 m/z,

563

an AGC target of 4e5, and a maximum injection rate of 50 ms in Profile mode. Features were

564

filtered for monoisotopic peaks with a charge state of 2-5 and a minimum intensity of 5e3, with

565

dynamic exclusion set to exclude features after 1 time for 60 seconds with a 10-ppm mass

566

tolerance. CID fragmentation was performed with collision energy of 35%, activation time of

567

10ms, and activation Q of 0.25 after quadrupole isolation of features using an isolation window

568

of 0.7 m/z, an AGC target of 1e4, and a maximum injection time of 35 ms. MS2 scans were then

569

acquired in the ion trap at rapid scan rate in Centroid mode. SPS-MS3 analysis was then

570

performed with a precursor selection range of 400-1600 m/z and precursor ion exclusion

571

tolerance of -50m/z to +5m/z. Up to 10 notches were selected using an MS2 isolation window

572

of 3 m/z for HCD fragmentation with a collision energy of 65%, which were analyzed in the

573

Orbitrap at 50k resolution with a scan range of 100-500, a maximum injection time of 500ms,

25

Downloaded from on May 14, 2021 at CALIFORNIA INST OF TECHNOLOGY

551