of 31
1
Bioorthogonal Non-Canonical Amino A
cid Tagging (BONCAT) enables
time-resolved
1
analysis of protein synthesi
s in native plant tissue
2
3
Weslee S. Glenn,
1
Shannon E. Stone,
1
Samuel H. Ho,
1
Michael J. Sweredoski,
2
Annie Moradian,
2
Sonja
4
Hess,
2
Julia Bailey-Serres
3
and David A. Tirrell
1
5
6
1 – Division of Chemistry and Chemical Engineering; California Institute of Technology; Pasadena, CA
7
91125
8
2 – Proteome Exploration Labora
tory; California Institute of Te
chnology; Pasadena, CA 91125
9
3 – Center for Plant Cell Biology; University of California at
Riverside; Riverside, CA 92521
10
11
Research Area
: Breakthrough Technologies
12
13
Summary
: Pulsing the non-canonical amino acid azidohomoalanine into
Arabidopsis thaliana
seedlings
14
enables in-gel visualization, physical enrichment, and identifi
cation of newly synthesized proteins.
15
16
List of Author Contributions
: W.S.G. conceived of and implem
ented the project, and performe
d most of
17
the experiments with contributi
ons from all other authors. S.H.
H., J.B-S and W.S.G. designed and
18
conducted gel imaging and imm
unoblotting experiments. A.M., S.H
. and W.S.G. designed and ran the mass
19
spectrometry experiments. S.E.S., M.J.S., S.H., J. B-S and W.S.
G. conducted all bioinformatic analyses of
20
the mass spectrometry data. S.E.S., M.J.S. and W.S.G. construc
ted figures with input from all other
21
authors. W.S.G, J.B-S and D.A.T. wrote the article with valuabl
e contributions from all authors.
22
Corresponding Author:
23
David A. Tirrell
24
California Institute of Technology
25
1200 E. California Blvd
26
MC 210-41
27
Pasadena, CA 91125
28
Telephone: (626) 395-3140
29
E-mail: tirrell@caltech.edu
30
31
Financial Sources
: This work was supported by the Gordon and Betty Moore Foundat
ion through grant
32
GBMF2809. W.S.G. was supported by a National Research Council Ford Foundation Post-Doctoral
33
Fellowship and a United Negro College Fund/ Merck Foundation Po
st-Doctoral Fellows
hip. The Proteome
34
Exploration Laboratory is supported by the Gordon and Betty Moo
re Foundation, through Grant
35
GBMF775, the Beckman Institute and the NIH through Grant 1S10RR
029594-01A1.
36
Plant Physiology Preview. Published on January 19, 2017, as DOI:10.1104/pp.16.01762
Copyright 2017 by the American Society of Plant Biologists
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
2
Abstract
37
Proteomic plasticity undergirds
stress responses in plants, and
understanding such
38
responses requires accurate measurement of the extent to which
proteins levels are
39
adjusted to counter external stim
uli. Here, we adapt bioorthogonal non-canonical amino
40
acid tagging (BONCAT) to interro
gate protein synthesis in veget
ative
Arabidopsis
41
thaliana
seedlings. BONCAT relies on the translational incorporation of
a non-canonical
42
amino acid (ncAA) probe into cellu
lar proteins. In this study,
the probe is the methionine
43
surrogate azidohomoalanine (Aha),
which carries a reactive azid
e moiety in its amino
44
acid side chain. The azide handle
in Aha can be selectively con
jugated to dyes and
45
functionalized beads to enable vi
sualization and enrichment of
newly synthesized
46
proteins. We show that BONCAT
is sensitive enough to detect
Arabidopsis
proteins
47
synthesized within a 30-min int
erval defined by an Aha pulse, and that the method can be
48
used to detect proteins made unde
r conditions of light stress,
osmotic shock, salt stress,
49
heat stress and recovery
from heat stress. We further establish
that BONCAT can be
50
coupled to tandem liquid chromat
ography-mass spectrometry (LC-M
S) to identify and
51
quantify proteins synthesized dur
ing heat stress and recovery f
rom heat stress. Our results
52
are consistent with a model i
n which, upon the onset of heat st
ress, translation is rapidly
53
reprogrammed to enhance the synthesis of stress mitigators and is again altered during
54
recovery. All experiments were
carried out with c
ommercially av
ailable reagents,
55
highlighting the accessibility of the BONCAT method to research
ers interested in stress
56
responses as well as translationa
l and post-transl
ational regul
ation in plants.
57
Introduction
58
Elevated temperatures, limited w
ater resources, and high salt concentrations in
59
arable soils are expected to pr
ofoundly and increasingly affect
the productiv
ity of crops
60
in the coming years (Mickelbart
et al
., 2015). Several prom
ising marker-assisted breeding
61
and genetic engineering strate
gies have been employed to help a
ddress this global
62
challenge (Kasuga
et al
., 1999; Lopes and Reynolds, 2010; Mickelbart
et al
., 2015). For
63
example,
HVA1
, a Late Embryogenesis Abundant
(LEA) gene from barley (
Hordeum
64
vulgare
L.) shown to delay leaf wilti
ng, was expressed in wheat (
Triticum aestivum
L.;
65
Sivimani
et al.
,
2000), rice (
Oryza sativa
L
.
; Xu
et al
., 1996)
and corn (
Zea mays
L
.
;
66
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
3
Nguyen and Sticklen, 2013) to produce
salt-tolerant lines with
higher water usage
67
efficiency. Despite these early and promising successes, target selection remains a critical
68
challenge associated with genetic
engineering and marker-assist
ed breeding (Bita and
69
Garets, 2013). Gaining further ins
ight into the physiological m
echanisms that govern
70
stress tolerance and adaptati
on will improve our ability to rat
ionally engineer crops.
71
Proteomic plasticity is a hallm
ark of the stress response in pl
ants, and was first
72
shown over 35 years ago by incorpor
ating radiolabeled amino aci
ds into proteins
73
synthesized during anaerobic st
ress in maize (Sachs et al., 198
0) and heat stress in
74
soybean (Key et al., 1981). With advances in mass spectrometry-based peptide
75
identification, new strategies have been developed to measure the extent to which protein
76
levels are adjusted in response
to environmental stimuli (Huot
et al
., 2014; Fristedt
et al.,
77
2015). But there are limitations t
o these technologies, which f
or the most part yield
78
information about steady-state
protein abundances. Moreover, tr
aditional shotgun
79
proteomics strategies provide reduced coverage in samples where
a few highly abundant
80
proteins predominate. For examp
le, RuBisCO comprises 30–60% of
the leaf proteome
81
and obstructs detection of less abundant proteins in leaf samples (Kim
et al.,
2013). To
82
counter this problem
specifically, Kim
et al.
developed a protamine sulfate precipitation
83
method to selectively deplete RuB
isCO and thus enrich for less
abundant proteins.
84
Pulsed stable isotope labeli
ng by amino acids in cell culture (pSILAC) was introduced to
85
monitor
de novo
protein synthesis, but poor label incorporation has prevented
it from
86
being widely adopted in plant systems. Other differential prote
omics techniques,
87
including SILAC (Lewandowska
et al
., 2013), hydroponic isotope labeling of entire
88
plants (HILEP; Bindschedler
et al.
2008),
13
CO
2
labeling (Chen
et al.
2011) and isobaric
89
tags for relative and absolute quantitation (iTRAQ; Ge
et al
., 2013, Pei
et al
. 2013), have
90
emerged in recent years as tools to probe
in vivo
protein synthesis in plants. Notably,
91
multiplexed quantifications with techniques such as iTRAQ are inherently problematic
92
due to known interference problem
s when using ion trap sources
(McAlister
et al
., 2014).
93
Modern instruments with MultiNotch MS3 capabilities can overcom
e these limitations,
94
but impose significant sensitivity penalties (McAlister
et al
., 2014). The instrumentation
95
required for such sophisticated e
xperiments is expensive and li
mits widespread use.
96
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
4
Identifying proteins that are tr
anslated within brief time inte
rvals remains
97
challenging because such proteins constitute a small fraction o
f the total proteome. While
98
transcriptomic approaches offer good time resolution (on the mi
nute time scale) (Kreps
et
99
al
., 2002; Preston
et al
., 2009), transcript abundances a
nd protein levels are frequent
ly
100
discordant (Vélez-Bermúd
ez and Schmidt, 2014; Fukao, 2015). Tra
nslating ribosome
101
affinity purification (TRAP) (Zanetti et al., 2005; Mustroph et al., 2009; Jiao and
102
Meyerowitz, 2010), followed by mRNA quantitation or ribosome fo
otprint mapping
103
(Juntawong
et al.
,
2014), provides proxies for protein synthesis at
specific time points in
104
environmental or developmental processes. However, measurements
of mRNA
105
association with ribosomes and
footprinting techniques may stil
l fail to provide accurate
106
information on changes in the proteome because of artifacts ari
sing from harvesting and
107
lysis steps (Ingolia, 2014) and t
urnover following protein synt
hesis. To complement the
108
existing chemical biology toolkit
for plant proteomics, we intr
oduce bioorthogonal non-
109
canonical amino acid tagging (BONCAT), which enables sensitive
detection and
110
identification of proteins synthe
sized within defined time inte
rvals.
111
In the BONCAT method, a non-canoni
cal amino acid is pulsed into
the cells of
112
interest, where it is incorporat
ed into newly synthesized prote
ins. Here we employed the
113
methionine surrogate azidohomoala
nine (Aha), which carries an azide moiety that is
114
amenable to bioorthogonal click c
hemistry (Figure 1). Bioorthog
onal click chemistry
115
refers to a set of reactions that are fast and highly selective
in complex biological settings,
116
and that can be carried out under
mild conditions (Sletten and
Bertozzi, 2009; McKay
117
and Finn, 2014). Pulse-labeling wit
h Aha allows the investigat
or to exploit the selectivity
118
of the copper-catalyzed
azide-alkyne cyclo
addition (CuAAC) and
the strain-promoted
119
azide–alkyne cycloaddition (SPAAC
) to visualize and enrich newl
y synthesized proteins
120
(Figure 1).
121
BONCAT has been used to study prot
eome-wide responses in mammal
ian and
122
microbial culture (Dieterich
et al.
,
2006; Zhang
et al
., 2014; Bagert
et al.
,
2015), and to
123
analyze sub-populations of cells in complex multicellular eukar
yotes (Erdmann
et al.
,
124
2015; Genheden
et al
., 2015; Yuet
et al.
, 2015), but has yet to be employed in plant
125
systems. Here, we applied BONCAT to
A. thaliana
to visualize proteins synthesized
126
within three hours after impositi
on of four significant abiotic
perturbations: light stress
127
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
5
(shift to darkness), high temperature, salt and osmotic stress.
We further coupled
128
BONCAT-assisted enrichment to tan
dem LC-MS to identify proteins synthesized
de novo
129
during transient heat stress and a recovery period. The results
from our screen are
130
consistent with a model in which cellular energy resources are rapidly channeled to
131
optimize a protective response at the onset of stress. We sugge
st that BONCAT should be
132
broadly useful in the investigation of plant physiology and dev
elopmental plasticity.
133
134
135
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
6
Results
136
BONCAT can be used to label newly synthesized proteins in
A. thaliana
seedlings
137
We pulsed Aha (1 mM) into 5-day-old seedlings (see methods) and
allowed
138
labeling to proceed for up to 3 h (Figure 2). To determine whet
her Aha is taken up and
139
incorporated into newly synthesized proteins, we first macerate
d aerial tissues of Aha-
140
exposed seedlings grown under non-
stress conditions with liquid
nitrogen and then
141
resuspended the resulting powder in a simple Tris buffer contai
ning SDS. Total protein
142
was isolated using chloroform-methanol precipitation, resuspended in a Tris buffer and
143
treated with TAMRA-alkyne (Fi
gure 2) under CuAAC conditions to
conjugate the
144
fluorophore to Aha-labeled protei
ns (Figure 1). We then resolve
d the proteins via one-
145
dimensional SDS-PAGE and measure
d in-gel fluorescence (Figure 2
B). Only Aha-
146
labeled proteins are tagged wit
h the fluorophore (Figure 2B and Figure 2C). We observed
147
light labeling in aer
ial tissues in periods as short as 30 min
(Figure 2B). These results
148
confirm that Aha was both transported into intact tissues and a
ctivated by the endogenous
149
methionyl-tRNA synthetase (MetRS).
150
To test the applicability of th
e BONCAT method to the study of
stress responses,
151
we used the Aha-labeling protocol described above, but also exp
osed the seedlings to a
152
short-term acute abiotic stres
ses for 3 h, including a shift to
darkness, osmotic shock (200
153
mM mannitol, a proxy for drought), high salt (300 mM NaCl) and
heat shock (37
°
C)
154
(Figure 3). Only the salt stress caused a change in phenotype,
where leaves collapsed
155
slightly upon exposure (Figure 3B). Importantly, the seedlings
exhibited no detectable
156
difference in phenotype resulting f
rom Aha exposure over the co
urse of the experiment
157
(Figure 3). We observed significant Aha labeling under all cond
itions in comparison with
158
negative controls, where seedli
ngs were grown under normal conditions but no Aha was
159
added (Figure 3C). Equal amounts of
total protein were loaded into each lane (Figure 3D)
160
to allow labeling levels to be
compared between conditions (Fig
ure 3C).
161
162
Aha incorporation is two-fold higher at an ambient temperature
of 37
°
C vs 22
°
C
163
While we observed protein labe
ling under each of the tested con
ditions (Figure
164
3C), we note that labeling inten
sity was approximately two-fold
stronger (2.1
±
0.4)
165
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
7
under heat shock versus other tested conditions, suggesting tha
t either translation rates
166
are faster or Aha uptake is gr
eater at higher temperatures. Se
veral studies with cell
167
suspension cultures and seedlings have shown decreases in prote
in synthesis rates at
168
elevated temperatures (Barnett
et al
., 1979 and Matsuura
et al.
, 2010). We note here that
169
although the ambient temperature in our experiment was 37
°
C, aerial tissues only
170
reached a peak temperature of approximately 31
°
C (measured by infrared thermometer
171
at the end of the stress period
just prior to harvest), likely
owing to evaporative cooling
172
effects. Notably, Ishihara
et al
. (2015) report higher protein synthesis rates at 28
°
C than
173
at 21
°
C in
Arabidopsis
seedlings using
13
CO
2
labeling. These results
are consistent with
174
ours, where the tissues experience a temperature ramp from 22
°
C to a maximum of 31
°
C.
175
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
8
Separately, to rule out the possibility that increased incorpor
ation of Aha was a
176
result of greater uptake, we performed pulses with Aha concentr
ations up to 4 mM and
177
labeled for 3 h at room temperature (SI Figure 1). Incorporatio
n was measured by in-gel
178
fluorescence. Under these conditions, Aha labeling appears to b
e saturated by 1 mM (SI
179
Figure 1). These results demonstrate that increased incorporati
on at higher temperatures
180
is likely not due to greater Ah
a uptake, but to increased prote
in synthesis.
181
BONCAT can be used to enrich newly synthesized proteins in
A. thaliana
seedlings
182
We tested our ability to enrich newly synthesized proteins through a
183
dibenzoazacyclooctyne (DBCO)-aga
rose bead pull-down method (Fig
ure 1 and Figure
184
2A). Specifically, we pulsed Aha into
A. thaliana
seedlings and allowed nascent proteins
185
to be labeled for 3 h at room te
mperature. To remove non-protein contaminants and
186
obtain better quality samples for mass spectrometry, we isolate
d total protein from
187
seedlings via trichloroacetic acid (TCA)-acetone precipitation
followed by phenol
188
extraction (Wang
et al
. 2006; Wu
et al
. 2014). We resuspended samples in phosphate-
189
buffered saline (PBS) supplemente
d with 1% SDS and checked for labeling by subjecting
190
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
9
total protein to TAMRA-alkyne conjugation (Figure 4B). A separa
te aliquot of total
191
protein was subjected to SPAAC conditions to conjugate the Aha-
labeled proteins to
192
DBCO-agarose beads. The resin was washed extensively to remove
non-specifically
193
bound proteins. Finally, Aha-labe
led proteins were eluted from the resin via trypsin
194
digestion and subjected to detergent removal and desalting prio
r to tandem mass
195
spectrometry (LC-MS/MS) (Orbitap Elite) and analysis via MaxQua
nt software.
196
To quantify enrichment, we performed parallel pull downs from seedlings at 22
197
o
C labeled with Aha for 3 h and s
eedlings unexposed to Aha. We subjected each pull
198
down to cleanup and tandem LC-MS and quantified the total sum o
f all
A. thaliana
MS1
199
peptide extracted-ion chromatogram (XIC) areas for both enriched and unlabeled mock
200
samples. We found enrichments of
at least 44-fold across three
biological replicates.
201
202
Identifying candidate proteins involved in thermotolerance and
recovery from heat
203
stress
204
After demonstrating the feasibility of labeling aerial tissues
under stress
205
conditions and developing an enrichment protocol, we sought to
identify the
de novo
206
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
10
synthesized proteins that were involved in heat stress and in r
ecovery from heat stress by
207
using a combination of BONCAT-assisted enrichment and LC-MS/MS.
We opted to
208
study heat stress because it has
been well characterized in
A. thaliana
at the level of
209
translation
(Matsuura
et al.
, 2010; Matsuura
et al
. 2012; Mittler
et al.
, 2012)
,
and because
210
thermotolerance has been shown to be induced by heating seedlin
gs to 38
°
C for 90 min
211
(Larkindale and Vie
rling, 2008), a timescale that is easily accessible to BONCAT
212
analysis (Figure 2B). Further, h
eat stress can be alleviated ea
sily to monitor protein
213
synthesis during the recovery from stress (McLoughlin
et al
. 2016), an aspect of stress
214
physiology that is relatively unders
tudied at the level of protein synthesis.
215
Seedlings were treated with A
ha and exposed to either heat shoc
k (37
°
C) or room
216
temperature control (22
°
C) conditions for 3 h. To study recovery, seedlings were first
217
exposed to ambient heat shock conditions (37
°
C) for 3 h, allowed to recover for 4 h at
218
room temperature (22
°
C), then treated with Aha at room temperature for 3 h.
219
Experiments were carried out in biological triplicate for each
condition (Figure 4A).
220
Aha-labeled proteins in experi
mental and control samples were c
onjugated to
221
DBCO-agarose beads and subjected to the enrichment protocol. In total, we identified
222
and quantified 3341 proteins across the four conditions [1 mM A
ha control at RT, 2 mM
223
Aha control at RT, heat stress (37
°
C) and recovery follo
wing heat stress (37
°
C to 22
224
°
C)] (Supplemental Table S1). We identified 2973 proteins acros
s the room temperature
225
control series 2 (2 mM Aha) and he
at shock samples alone (Figur
e 4C).
226
To assess enrichment of known heat-responsive proteins, we firs
t filtered our
227
dataset for proteins that were
either significantly upregulated
in response to heat (p-value
228
< 0.01) or that were found in all t
hree heat-shock replicates a
nd in none of the control
229
series 2 (2 mM
Aha) replicates [
Table S1; Label-Free Quantitation (LFQ) value = 0,
230
Columns DV - DX). These crite
ria identified a total of 189 pro
teins as heat-responsive
231
BONCAT-enriched proteins (Table
S1; Column EL). Proteins with a gene ontogeny
232
(GO) annotation of ‘Response to H
eat’ were found to be signific
antly over-represented
233
(p-value 8x10
-22
; Fisher’s exact test) in the population of heat-re
sponsive BON
CAT-
234
enriched proteins. These results
clearly demonstrate that the B
ONCAT method detects
235
enrichment of
de novo
synthesized ‘heat-responsive
proteins’ in response to 3 h of heat
236
stress. We note that this assay detected proteins encoded by nu
clear, mitochondrial and
237
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
11
plastid genes, showing that Aha
is incorporate
d into proteins s
ynthesized in different
238
cellular compartments.
239
We constructed a volcano plot of proteins shared between heat s
hock samples and
240
control series 2 samples to visua
lize proteins with statistical
ly significant fold-changes
241
(Figure 4D). Our lis
t of up-regulated proteins contains many kn
own heat stress markers,
242
including ClpB1, Hsp90-1, probable
mediator of RNA polymerase I
I transcription
243
subunit 37c and HSP70-5 (Queitsch
et al
., 2000; Lin
et al.
,
2001; Sung
et al
., 2001;
244
Takahashi
et al
., 2003; Yamada and Nishimura, 2008)
. Our analysis also identi
fied
245
proteins with statistically si
gnificant fold-changes that have
not been annotated
246
previously.
247
We performed principal component
analysis on the basis of normalized LFQ
248
values for each protein (Figure 5). We found three
distinct clu
sters: control samples, heat
249
shock samples and recovery samples. These results illustrate th
e repeatability of
250
biological replicates in BONCAT a
nalysis, and the clustering of
the two control series (1
251
mM and 2 mM Aha) suggests that Aha does not cause significant p
erturbation of protein
252
synthesis at these concentrations. Furthermore, the separate clustering of the control and
253
recovery samples shows that metabolism does not simply return t
o the pre-imposition
254
state following heat stress.
255
Next, we constructed heat maps t
o compare protein levels across
conditions
256
(Figure 6; Supplemental Table S1; Supplemental Figure S2). This analysis demonstrates
257
the distinction in BONCAT-labeled
proteins under the three conditions, including the
258
marked upregulation of heat respons
e proteins under heat shock.
Notably, many
259
BONCAT-labeled proteins highly expressed during heat shock are
synthesized at lower
260
levels during the recovery period than under control conditions
, clearly demonstrating
261
that seedlings rapidly adjust
to changing conditions in part by
altering the synthesis of
262
proteins.
263
To validate our BONCAT results, w
e performe
d immunoblot detecti
on of two up-
264
regulated proteins: ClpB1 (HSP101) and HSP 70-5. For this purpose, 5-d-old seedlings
265
were grown identically to those
in the BONCAT screen, then exposed to 22
°
C for 3 h
266
(control), 37
°
C for 3 h (heat shock), or 37
°
C for 3 h then 22
°
C for 7 h (recovery; these
267
conditions mimic the 4-h ‘rest’
period plus the 3-h labeling period in the BONCAT
268
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
12
experiment). We then extracted t
otal protein in a procedure ide
ntical to protein extraction
269
for analysis by LC-MS/MS. As anticipated, we observed strong in
duction for both ClpB1
270
and HSP 70-5 under heat stress (Figure 7). Importantly, immunob
lotting detected
271
differences in abundance across treatment samples, irrespective
of time of synthesis. In
272
contrast, BONCAT measures protein
synthesized within specified
time frames.
273
Most of the proteins highly upreg
ulated during heat shock are downregulated
274
during the recovery period according to BONCAT detection (Figur
e 6; Figure S2). We
275
hypothesized that these highly upr
egulated proteins would not be degraded rapidly during
276
the recovery period because ther
mal priming, a thermotolerance mechanism, has been
277
shown to occur under similar heat stress conditions on similar
timescales (Larkindale and
278
Vierliing, 2008; Mittler
et al
., 2012). The results of our immunoblotting support the
279
hypothesis. Specifically, ClpB1 and HSP70-5 were not degraded during the recovery
280
from heat shock (Figure 7), although their
de novo
synthesis was significantly reduced
281
(Table S1). Therefore, ClpB1 and HSP70-5 are stable over the co
urse of our experiment.
282
283
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
13
BONCAT as a hypothesis generator
284
Gratifyingly, the BONCAT screen revealed 80 validated biomarker
s (GO
285
molecular function ‘response to h
eat’) of the heat stress respo
nse in
A. thaliana
,
including
286
ClpB1 (HSP101) and HSP70-5. Notably, other proteins without an
annotation of
287
‘response to heat’ were also upreg
ulated during heat stress, su
ggesting a possible role in
288
thermotolerance. Measuring proteins synthesized during recovery
from heat stress
289
yielded new information. For example, during recovery we observ
ed induction (2.75-
290
fold) of zeaxanthin epoxidase (ZEP), which catalyzes the first
committed step in the
291
biosynthesis of abscisic acid (Xiong and Zhu, 2003), a hormone
known to promote
292
stomatal closure (Morillo and Chrispeels, 2001; Park
et al.
2015). Stomatal closure, in
293
turn, minimizes water loss due to
evaporative cooling (Xiong and Zhu, 2003). Therefore,
294
we hypothesize that under these
conditions ZEP is downregulated during heat stress to
295
promote evaporative cooling, the
n upregulated during recovery t
o promote stomatal
296
closure and prevent desiccation
. While detailed assignment of f
unctional roles to heat-
297
responsive proteins in the context of imposition and recovery i
s beyond the scope of this
298
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
14
study, the observed changes in p
rotein synthesis suggest mechan
istic hypotheses that
299
merit further evaluation.
300
301
302
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
15
Discussion
303
Proteins are cellular workhorses that carry out tightly orchestrated developmental
304
and adaptive programs. As sessile organisms, plants must retain
a high degree of
305
proteomic plasticity to enable
rapid responses to a barrage of environmental cues (Huot
et
306
al
., 2014; Fristedt
et al.
,
2015; Mickelbart
et al
., 2015). In plants, as in most eukaryotes,
307
mRNA abundance is often a poor pr
oxy for protein levels (Ingoli
a, 2014; Velez-
308
Bermudez and Schmidt, 2014; Fukao, 2015), as indicated by the d
iscordance between
309
total mRNA abundance and polyribosome-associated mRNAs under co
ntrol and abiotic
310
stress conditions (reviewed by Roy and von Arnim, 2013). To add
ress this issue, we used
311
BONCAT in intact
A. thaliana
seedlings to identify and quan
tify proteins synthesized
312
under conditions of abiotic stre
ss. As implemented here, BONCAT
enables identification
313
of proteins synthesized over periods of a few hours, and requir
es no manipulation of the
314
translational machinery or pres
umption that a ribosome-associat
ed mRNA will produce a
315
stable protein (Figure 2). Furthermore, endogenous methionine l
evels do not need to be
316
depleted for Aha to be incorpor
ated into newly synthesized proteins.
317
In an initial test of the met
hod, we found evidence that newly synthesized proteins
318
can be labeled in seedlings subjected to a variety of abiotic s
tresses, including osmotic
319
shock, high salt and heat shock (Figure 3). We then compared th
e populations of
A.
320
thaliana
proteins made under normal growth conditions to those made und
er conditions
321
of mild heat stress and during rec
overy from heat stress (Figur
e 4). Changes in protein
322
synthesis in response to heat s
tress were readily observed in p
eriods as short as 3 h. We
323
used bioinformatic (Figure 4; F
igure 5 and Figure 6) and immuno
blotting analyses
324
(Figure 7; Figure S3 and Figure
S4) to validate the BONCAT-labe
led proteins under
325
three conditions. Unsurprisingl
y, the GO annotation ‘Response
to Heat’ was over-
326
represented among proteins upregul
ated in response to heat shoc
k. This result is
327
congruent with previous studies
showing selective induction of
stress m
itigators during
328
mild heat stress
.
Our identification of
de novo
synthesized markers of the heat shock
329
response provides validation of t
he BONCAT method as a tool for
the study of proteome
330
dynamics in plants. At the same time, we found many proteins, including presumably
331
low abundance transcription factors like BIM1, that have not previously been associated
332
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
16
with mitigation of heat stress, illustrating the potential valu
e of time-resolved proteomic
333
studies as a source of new
mechanistic hypotheses.
334
Furthermore, we assessed protei
n synthesis during recovery from
heat stress and
335
found proteins most highly upregulated in response to heat stre
ss were synthesized during
336
recovery at levels similar to or
lower than pre-heat shock levels (Figure 6). This result
337
showcases a mechanism by which pl
ants avoid synthesizing a surfeit of stress-associated
338
proteins during rapidly changi
ng conditions. Based on reduced
de novo
synthesis and
339
maintained abundance, neither ClpB1 nor HSP 70-5 was rapidly de
graded during the
340
recovery period, consistent with pr
evious work showing that plants can develop “thermal
341
memory” (Larkindale and Vierl
ing, 2008) that protects seedlings exposed to mild heat
342
stress from subsequent assaul
ts. This observation suggests that the BONCAT method
343
may also be applied to consider dyn
amics in protein turnover, a
s the protein detected will
344
reflect both the synthesis and turnover within the time-period
of labeling. Importantly,
345
turnover and synthesis may be di
fferentially regulated. Thus, t
he BONCAT method is
346
unlikely to detect highly
unstable proteins.
347
It is worth noting that our mode
l employed vegetative seedlings
. While
348
understanding stress responses at
all developmental stages is i
mportant to the engineering
349
of more robust crops, previous stu
dies have shown that yield lo
sses are caused most
350
prominently by assaults to reproductive tissues (Young
et al
., 2004; Zinn
et al
., 2010).
351
The BONCAT method should prove usef
ul across many developmental stages and tissues
352
and under a wide variety of conditions, and can help unravel th
e genetic basis of traits
353
involved in tolerance to both abiotic and biotic stresses. Nota
bly, all reagents and probes
354
used in this study are commercially available, further undersco
ring the ease by which this
355
methodology can be readily adopted
by laboratories with access
to proteomic facilities.
356
357
358
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
17
Methods
359
360
Plant Growth Conditions
361
362
Arabidopsis thaliana
seeds (Col-0 accession) were subjected to vapor phase
363
sterilization (100 mL of 6% (v
/v) commercial bleach and 3 mL of
37% HCl (v/v)) for
364
approximately 3 h. Seedlings were then dispensed on 40 mL soli
d media (100 x15 mm
365
petri dishes) containing 0.5x str
ength Murishige Skoog (MS) sal
ts (Sigma-Aldrich; St.
366
Louis, MO), 1x MS vitamins (Si
gma-Aldrich; St. Louis, MO), 0.3%
(w/v) sucrose and
367
0.9% (w/v) Phytagel (Sigma-Aldric
h; St. Louis, MO). Plates were
inverted at 4
°
C in
368
darkness for 2 d to break seed dormancy. Plates were then place
d at 22
°
C under 24-h
369
daylight and positioned so that roots would grow into the mediu
m. After germination,
370
plants were allowed to grow for 5 d under constant light.
371
372
Seedling flood technique to del
iver non-canonical amino acid
373
The protocol was adapted from the study of Ishiga
et al.
(2011). Aha labeling
374
medium containing 0.5x Murishige
Skoog (MS) salts (Sigma-Aldric
h; St. Louis, MO),
375
5% sucrose, 0.0025% Silwet L-77 (
Lehle Seeds, Round Rock, TX),
1 mM
376
azidohomoalanine (Aha, Iris Biotech
Gmbh; Marktredwitz, Germany) in 10 mM citrate
377
buffer (pH 5.6) was freshly prep
ared prior to running the exper
iment. Seedlings were
378
fully submerged in the Aha labeling medium for 2 min. After the
medium was decanted,
379
the petri dish covers were repl
aced and the plates were wrapped
in foil. The seedlings
380
were then exposed to short-term
acute abiotic stresses for 3 h.
Shift to darkness was
381
accomplished by covering plates with foil. To approximate osmot
ic shock, Aha labeling
382
medium was supplemented with 200
mM mannitol and pulsed onto se
edlings during
383
delivery of Aha. High salt stress was accomplished by supplemen
ting Aha labeling
384
medium with 300 mM NaCl. To heat
shock samples, plates were pl
aced in an oven at an
385
ambient temperature of 37
°
C immediately following the 2-min submersion in Aha
386
labeling me
dium. All proteomi
cs experiments were initiated on
the morning of day 5
387
(germination is defined as day 0).
388
389
In-gel fluorescence to mon
itor protein labeling
390
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
18
After labeling with Aha, aerial
tissues were immediately harves
ted, flash frozen with
391
liquid nitrogen and stored at -80
°
C until subsequent workup steps. Frozen tissues were
392
macerated in liquid nitrogen. Fro
zen powder was immediately transferred into an
393
Eppendorf tube containing 1 mL of
lysis buffer (100 mM Tris, pH 8 containing 4% w/v
394
SDS). Lysates were subjected to sonication at 80
°
C for 40 min, then subjected to further
395
heating at 95
°
C for 30 min. Cellular debris was removed by centrifugation at
16,100
396
RCF for 5 min. The conditions for labeling the lysate by CuAAC
for in-gel fluorescence
397
were based on a previousl
y published protocol (Hong
et al
., 2009). Protein concentrations
398
in the cleared lysate were meas
ured via bicinchoninic acid (BCA
) assay. Forty
399
micrograms of protein lysate was added to phosphate-buffered saline to a final total
400
volume of 221
μ
L. In a separate tube, the al
kynyl dye, copper (II
) sulfate and tris(3-
401
hydroxypropyltriazolylmethyl)amine
(THPTA) were premixed and allowed to react for 3
402
min in the dark. Then, aminoguanidi
ne HCl and sodium ascorbate
were added to the
403
lysate-PBS mixture. Final concen
trations were as follows: alkynyl dye (2.5
μ
M), CuSO
4
404
(1 mM), THPTA (5 mM), aminoguanidi
ne HCl (5 mM), sodium ascorba
te (5 mM). All
405
components were gently mixed (one
inversion) and allowed to react for 15 min in the
406
dark at room temperature without
shaking. Proteins were then extracted by
407
methanol/chloroform/water preci
pitation. Pellets were washed ex
tensively (at least three
408
times). The pellet was then disso
lved in 2X SDS sample loading
buffer and sonicated for
409
30 min at 80
°
C. Samples were heated to 95
°
C for 5 min and electrophoresed in pre-cast
410
4-12% Bis-Tris polyacrylamide ge
ls. The gel was washed with a fixing solution (50%
411
water, 40% ethanol, 10% acetic a
cid) in the dark for 10 min and
rinsed twice with water
412
(2 × 10 min) prior to being subjected to fluorescence imaging with an excitation laser at
413
532 nm and an emission band pass fi
lter at 580 nm (Typhoon GE H
ealthcare). After
414
fluorescence imagining, the gel wa
s stained with Colloidal Blue
(InstantBlue, Expedeon,
415
San Diego, CA) for 1 h and imaged to e
nsure equal protein loading among all lanes. To
416
obtain relative fluorescence va
lues, signal intensity was measured for the entire lane in
417
both the TAMRA (fluorescence) cha
nnel and the colloidal blue ch
annel. Next, the ratio
418
of fluorescence intensity to col
loidal blue intensity was calcu
lated and normalized to the
419
control lane. Calculating the re
lative fluorescence values as r
atios in this manner allows
420
for comparison between lanes even
if lane loading varies slight
ly.
421
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
19
Cell lysis and protein extraction for enrichment
422
Samples were prepared in biological triplicate with 1 plate (ap
proximately 0.5 -
423
0.7 g tissue fresh weight) represen
ting one biological replicat
e. Samples were prepared
424
with protocols adapted from Wang
et al
. (2006) and Wu
et al
. (2014). Frozen aerial
425
tissues were macerated in liquid nitrogen with a mortar and pes
tle. The resultant powder
426
(0.1-0.3g) was transferred into a 1.7
-mL tube (typically tissues from 1 plate were divided
427
into 6 separate 1.7-mL tubes). The
pellets were washed with 10%
trichloroacetic acid
428
(TCA) in acetone by filling the tubes, mixing them well by vort
exing then centrifuging at
429
16,000 RCF for 3 min at 4
°
C. The supernatant was removed. The pellet was washed with
430
80% methanol containing 0.1 M ammonium acetate then with 80% ac
etone. The pellets
431
were allowed to dry at room temperature for 15 min to remove re
sidual acetone. Next, 0.8
432
mL of liquefied phenol (pH 8.0, Sigma-Aldrich, catalog number P
4557) and 0.8 mL of
433
SDS dense buffer [30% sucrose, 2% SDS, 0.15% sodium azide (to prevent bacterial
434
growth), 0.1 M Tris-HCl, pH 8.0 a
nd 2-mercaptoethanol to a fina
l concentration of 5%
435
added fresh] were added to the t
ube. The contents were mixed th
oroughly and incubated
436
for 5 min at room temperature in a fume hood. Next the tubes we
re centrifuged at 16,000
437
RCF for 3 min at 4
°
C. The upper phenol phase was tra
nsferred into a new 1.7-mL tub
e
438
while taking precautions not to disturb the middle SDS interfac
e. The new 1.7-mL
439
Eppendorf tube was filled wit
h methanol containing 0.1 M ammonium acetate and stored
440
at -20
°
C overnight to precipitate the
protein. The tubes were then cen
trifuged at 16,000
441
RCF for 5 min at 4
°
C. The pellet was washed once w
ith 100% methanol, then with 80%
442
acetone. Next, the protein pellets were resuspended in phosphat
e-buffered saline [PBS,
443
pH 7.4, 1% SDS, 100 mM chloroacetami
de and 1X cOmplete EDTA-free protease
444
inhibitors (Sigma Aldrich, Catalog Number
11873580001)] then pooled. Protein
445
concentrations were measured via BCA assay. Approximately 0.3 m
g total protein was
446
carried through the enrichment procedure.
447
448
Enrichment Procedure
and LC-MS/MS Sample Preparation
449
Samples were adjusted to a c
onsistent volume (approximately 500
μ
L) with 1%
450
SDS in PBS. Then, samples were
diluted 2X with 8 M urea contain
ing 1X EDTA-free
451
protease inhibitor (Sigma
Aldrich, Catalog Number
11873580001) and 0.85 M NaCl.
452
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.
20
Separately, DBCO-agarose beads
(50 uL of 2X slurry
; Click Chemi
stry Tools,
453
Catalog Number 1034-2) were wa
shed 3X with 0.8% SDS in PBS. The
n, protein samples
454
were added to the resin and shaken at 1200 rpm at room temperat
ure overnight (for at
455
least 16 h). The resin was was
hed with 1 mL water. Then, 0.5 m
L dithiothreitol (1 mM
456
in 0.8% SDS in PBS) was added to
the resin and incubated with shaking (1200 rpm) for
457
15 min at 70
°
C. The supernatant was removed, and free thiols were blocked w
ith
458
iodoacetamide (0.5 mL of a solution at 7.4 mg/mL dissolved in P
BS with 0.8% SDS) for
459
30 min in the dark at 1200 rpm.
460
The resin was transferred to a spin column (Poly-prep ® chromat
ography
461
columns, Bio-Rad, catalog numb
er 731-1550) and subjected to the
following washes: 8 x
462
5 mL 0.8% SDS in PBS, 8 x 5 mL 8
M urea, 8 x 5 mL 20% acetonitr
ile (ACN). For the
463
second wash with each solution, the column was capped and allow
ed to sit for 10 min
464
prior to draining. After the was
hes, beads were transferred to
an Eppendorf tube with
465
10% ACN in 50 mM ammonium bicarbon
ate. Tubes were centrifuged
for 5 min at 2000
466
g and the liquid was removed. Then, 100
μ
L 10% ACN in 50 mM ammonium
467
bicarbonate was added to the b
eads and 100 ng trypsin was added
. The beads were
468
shaken at 1200 rpm and 37
°
C overnight. The supernatant w
as collected, and beads were
469
washed twice with 150
μ
L 20% ACN. Supernatants from the 20% ACN washes were
470
pooled with the supernatant from the overnight tryptic digest.
Pooled supernatants were
471
filtered (Pierce centrifuge colu
mns 0.8 ml, ThermoFisher Scient
ific, catalog number
472
89868) to remove any beads that
carried through, then dried on
a speedvac.
473
Digested peptides w
ere redissolved in 100
μ
L 50 mM ammonium bicarbonate and
474
treated with HiPPR detergent removal resins (ThermoFisher Scien
tific, catalog number
475
88306). Finally, digested peptides were subjected to desalting
clean-up step with a ZipTip
476
(C18).
477
478
LC-MS/MS analysis
479
Trypsin-digested samples were s
ubjected to LC-MS/MS analysis on a nanoflow
480
LC system,
EASY-nLC II, (Thermo
Fisher Scientific) coupled to a Orbitrap Elite Hybrid
481
Ion Trap-Orbitrap Mass Spectrometer (Thermo Fisher Scientific,
Bremen, Germany)
482
www.plantphysiol.org
on January 23, 2017 - Published by
www.plantphysiol.org
Downloaded from
Copyright © 2017 American Society of Plant Biologists. All rights reserved.