Plant Physiology Preview. Published on September 6, 2016, as DOI:10.1104/pp.16.01082
1
Short title: Gene expression in pine heartwood transition zone
2 3
Corresponding author:
4
Teemu H. Teeri
5
Department of Agricultural Sciences
6
University of Helsinki
7
Latokartanonkaari 7
8
00790 Viikki, Finland
9
Phone: +358 2941 58380
10
Email:
[email protected]
11 12
Research areas:
Biochemistry and Metabolism
13
Secondary areas: Genes, Development and Evolution
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
1 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
Copyright 2016 by the American Society of Plant Biologists
29
Developmental changes in Scots pine transcriptome during heartwood
30
formation
31 32
Kean-Jin Lim1, Tanja Paasela1, Anni Harju2, Martti Venäläinen2, Lars Paulin3,
33
Petri Auvinen3, Katri Kärkkäinen4 and Teemu H. Teeri1*
34 35
1
36
00014, University of Helsinki, Finland
37
2
38
Punkaharju, Finland
39
3
Institute of Biotechnology, P.O. Box 56, 00014, University of Helsinki, Finland
40
4
Natural Resources Institute Finland (Luke), P.O. Box 413, 90014 University of
41
Oulu. Finland
Department of Agricultural Sciences, Viikki Plant Science Centre, P.O. Box 27, Natural Resources Institute Finland (Luke), Finlandiantie 18, 58450
42 43
One-sentence Summary
44
The Scots pine heartwood formation takes place in the summer and is marked
45
by programmed cell death.
46 47
Footnote:
48
List of author contributions: THT, KK, AH and MV conceived and designed
49
the research; AH and MV obtained the plant material; LP and PA were
50
responsible for designing the library phase and conducting SOLiD sequencing;
51
KJL and TP designed, planned and carried out the experiments; KJL performed
52
data analysis; THT contributed in interpreting the data; KJL and TP wrote and
53
THT revised the manuscript, which all authors read, commented and approved.
54 55
Funding information: This work was funded by the National Technology
56
Agency of Finland (TEKES).
57 58
* Corresponding author: Teemu H. Teeri (
[email protected])
59
2 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
60
Abstract
61 62
Scots pine (Pinus sylvestris L.) wood is desired in woodworking industries due
63
to its favourable timber characteristics and natural durability that is contributed
64
by heartwood extractives. It has been discussed whether the Scots pine
65
heartwood extractives (mainly stilbenes and resin acids) are synthesized in the
66
cells of the transition zone between sapwood and heartwood, or if they are
67
transported from the sapwood. Timing of heartwood formation during the yearly
68
cycle has also not been unambiguously defined. We measured steady state
69
mRNA levels in Scots pine transition zone and sapwood using RNA
70
sequencing. Year round expression profiles of selected transcripts were further
71
investigated by quantitative RT-PCR. Differentially accumulating transcripts
72
suggest that, of the Scots pine heartwood extractives, stilbenes are synthesized
73
in situ in the transition zone and gain their carbon-skeletons from sucrose and
74
triglycerides. Resin acids, on the other hand, are synthesized early in the spring
75
mainly in the sapwood, meaning that they must be transported to the heartwood
76
transition zone. Heartwood formation is marked by programmed cell death that
77
occurs during the summer months in the transition zone.
78 79 80 81 82 83 84 85 86 87 88 89
3 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
90
Introduction
91 92
Wood biosynthesis is a complex process that involves several developmental
93
activities such as cell division, cell expansion, cell wall thickening, and
94
programmed cell death (Plomion et al., 2001). In many, but not all tree species,
95
the final stage in wood development is formation of heartwood (HW). HW is
96
traditionally defined as “the inner layers of wood, which, in the growing tree,
97
have ceased to contain living cells, and in which the reserve materials (e.g.
98
starch) have been removed or converted into heartwood substances”, while
99
sapwood (SW) is defined as “the portion of the wood that in the living tree
100
contains living cells and reserve materials” (IAWA committee, 1964). Heartwood
101
can also be defined as the region where extractives have accumulated and by
102
this definition it can be metabolically active (Beekwilder et al., 2014; Celedon et
103
al., 2016) although all changes associated with heartwood formation may not
104
have yet occurred (Taylor et al., 2002).
105 106
The biological role of HW is still unclear, but it may be important in long-term
107
resistance against pathogens (Venäläinen et al., 2004). Nevertheless, HW has
108
a clear economic importance due to natural decay resistance of HW timber in
109
many tree species (Singh and Singh, 2011) and due to increasing
110
pharmaceutical and biocide applications of heartwood substances (Kampe and
111
Magel, 2013).
112 113
Heartwood can be found in both angiosperm (e.g. Acacia, Catalpa, Juglans and
114
Robinia) and gymnosperm (e.g. Cryptomeria, Larix, Picea and Pinus) tree
115
species. It is physiologically inactive, drier compared with the SW and filled with
116
HW extractives (Taylor et al., 2002). SW, on the other hand, is physiologically
117
active, has high moisture content, and functions in water, nutrient and sugar
118
transport between roots and the foliage (Taylor et al., 2002).
119 120
As the tree grows, SW is converted to HW. This takes place in the narrow
121
transition zone (TZ), the living cells in which reserve materials like starch are
4 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
122
consumed. TZ is drier than SW, but not as dry as HW (Bergström, 2003; Hillis,
123
1987). The width of the TZ varies both within and between tree species and
124
could be influenced by seasonal as well as other environmental aspects. In
125
general, the TZ is about one to three annual rings wide (Hillis, 1987), in Scots
126
pine usually one to two annual rings wide (Gustafsson, 2001).
127 128
It is still unclear when and how the initiation of HW formation occurs.
129
Involvement of the plant hormones ethylene and auxin in regulation of the HW
130
formation have been suggested (Nilsson et al., 2002; Shain and Hillis, 1973;
131
Hillis, 1968). Spicer (2005) described HW formation as a form of programmed
132
cell death (PCD) where the organelles of ray parenchyma cells in the TZ are
133
gradually disintegrated. Several studies have shown that the cell wall structure
134
of ray parenchyma cells change and the cells gradually die towards HW at the
135
boundary between SW and HW (Yang et al., 2003; Nakaba et al., 2008; Magel
136
et al., 2001; Nakada and Fukatsu, 2012; von Arx et al., 2015). In addition to
137
plant hormones, Nakada and Fukatsu (2012) suggested that desiccation at the
138
boundary between SW and HW initiates HW formation.
139 140
The specific timing of the HW formation is unclear and may certainly be species
141
specific. Different studies have described the process taking place at different
142
times of the year. HW formation in black locust (Robinia pseudoacacia) took
143
place in autumn, while in radiata pine (Pinus radiata) and black walnut (Juglans
144
nigra) it occurred during the dormancy period in winter (Taylor et al., 2002;
145
Kampe and Magel, 2013). Bergstöm et al. (1999) concluded that there is no
146
specific time of the year for HW formation in Scots pine. Indeed, the timing of
147
HW formation is not clearly understood nor has been satisfactorily described up
148
to date.
149 150
Two types of HW formation were distinguished by Magel (2000). In type I, or
151
Robinia-type of HW formation, HW extractives are biosynthesised in situ and
152
accumulate in the TZ. Precursors to the extractives are not present in the
153
ageing SW. In type II, or Juglans-type of HW formation, HW extractives are
5 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
154
formed through transformation of phenolic precursors by hydrolysis, oxidation
155
and polymerization in the TZ. In type II HW formation, HW extractive precursors
156
are gradually accumulating in the ageing SW (Magel, 2000; Kampe and Magel,
157
2013). Deposition of HW extractives and PCD marks the end of the HW
158
formation (Spicer, 2005). Many HW extractives are synthesised also in living
159
tissues when trees are under biotic or abiotic stress and thus contribute also to
160
the active defence of SW and other living tissues (Chiron et al., 2000b).
161 162
In Scots pine, HW extractives consist of the stilbenes pinosylvin (PS) and
163
pinosylvin monomethyl ether (PSME), resin acids and free fatty acids
164
(Saranpää and Nyberg, 1987). Variation of Scots pine HW phenolic extractive
165
content between tree individuals is wide and strongly correlates with decay
166
resistance of HW (Harju and Venäläinen, 2006; Leinonen et al., 2008). Although
167
extractive content is influenced by the environment (Magel, 2000), it is highly
168
inherited (Fries et al., 2000; Partanen et al., 2011).
169 170
It is still unclear if Scots pine HW extractives are synthesized in the cells of TZ
171
between SW and HW, or if they are transported from the SW. Furthermore, the
172
timing of HW formation during the yearly cycle has not been unambiguously
173
defined. In this work we used a transcriptomic approach to characterise
174
changes in gene expression during HW formation, i.e. in TZ compared to SW.
175
Using a set of indicator genes, defined through transcriptomics, we further
176
addressed the question of timing of HW formation by quantitating gene
177
expression in SW and TZ throughout the year in two Scots pine individuals.
178 179
6 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
180
Results
181
Localisation of the transition zone and transcript sampling
182
Increment cores were pencil marked in the field for the border between the wet
183
and the dry zone and photographed under UV illumination in the laboratory to
184
localise the fluorescing stilbenes. Fluorescence was observed in the whole of
185
the dry region, except for the outermost about one annual ring wide zone
186
(Figure 1A).
187 188
In order to determine the TZ in molecular terms, increment cores from two
189
trees, collected in June 2010, were sectioned per year ring around the pencil
190
mark (Figure 1A) and total RNA was extracted. The RNA yield, based on
191
absorbance at 260 nm, was comparable from zones 5-8 and somewhat less
192
from zone 4 (Figure 1B and Table S1). Zones 1-3 yielded much less material
193
absorbing at 260 nm, except for the June sample of tree 11 (Table S1). These
194
samples were analysed using Bioanalyzer and RT-PCR with primers designed
195
for a Scots pine actin gene. Results indicated that the absorbance was not due
196
to RNA (Figure S1), in line with the observations by Nakaba and colleagues
197
(2008) and with the definition of HW (IAWA committee, 1964). Semiquantitative
198
RT-PCR for the pine stilbene synthase (STS) encoding transcript was
199
performed for each sample. The outermost dry zone (zone 4) showed strongest
200
STS expression in all sampled series (Figure S2) and was designated as the
201
TZ, well in accordance of the definition by Bergström (2003) and Hillis (1987).
202
The SW sampling position was specified as 4 or 5 annual rings outwards from
203
the TZ (Figure 1A).
204 205
To choose an ideal sampling time, we harvested increment cores from two trees
206
during the summer months June, July and August of 2010. Semiquantitative
207
RT- PCR for the STS transcript was carried out from SW and TZ (Figure S2).
208
The pattern was consistent through these months, showing STS activity in TZ
209
but not in SW. As the earliest sampling month (June) tended to show stronger
210
activity than samples towards the end of summer, it was chosen as the full
211
scale transcriptome sampling month for the following year.
7 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
212 213
Scots pine transcriptomes
214
About thirty increment cores (per tree) were drilled from four 46 years old
215
unrelated Scots pine trees at breast height at the Leppävirta progeny trial stand
216
in June 2011. SW and TZ samples from each tree were pooled for RNA
217
extraction, tested for STS expression (Figure S3) and sequenced using the
218
SOLiD platform. 12-17 million paired-end reads were obtained from each RNA
219
sample. All transcriptome libraries were mapped against the Pinus EST
220
collection version 9.0 (The Gene Index Databases, 2014), the mapping rate
221
being 64-66 % for each library (Table S2). The Pinus EST collection is a
8 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
222
combined collection of expressed sequence tags from 36 pine species (P. taeda,
223
P. sylvestris, P. radiata etc.). Sequences from different species show high
224
similarity and their assembly into “tentative consensus” (TC) sequences has
225
been done as if they originated from a single species (Quackenbush et al.,
226
1999). Therefore, the collection of 77 326 TC sequences in the Pinus EST
227
collection is not an indicator of the pine gene number in any sense, nor do they
228
represent necessarily assemblies that are species specific, although many
229
assemblies match one-to-one with cDNA molecules isolated from individual
230
species. For example, while TC154538 is 98 % identical to the Scots pine STS
231
mRNA sequence S50350, there are altogether 25 STS encoding TCs matching
232
S50350 for 10-100 % in length and 88-98 % for sequence identity. By the
233
nature of the Pinus EST collection, the number of similar TCs sharing
234
differential expression has no meaning (only their annotation does), although
235
they may sometimes represent gene family members in Scots pine.
236 237
We also mapped the reads to an annotated Trinity (Haas et al., 2013) assembly
238
generated from the RNA sequencing (RNA-Seq) reads themselves after
239
converting them first to nucleotides. While the mapping rate was often higher
240
and fold changes for differentially expressed transcripts in some cases larger,
241
the expression data obtained was very similar to that obtained with Pinus EST
242
collection as reference. We present here results with the Pinus EST collection
243
since it includes also transcripts that are not expressed in our samples (e.g.,
244
transcripts for resin acid biosynthesis, see below) and the assemblies (TCs) are
245
longer. Although the full genome of Pinus taeda was recently published with
246
draft gene models (Wegrzyn et al., 2014), it performed very poorly as a target
247
for mapping our reads and the mapping rate was only approximately 6 % for
248
each library (Table S2).
249
Differential expression analysis was carried out using edgeR (see materials and
250
methods). A total of 1673 transcripts had statistically significant (false discovery
251
rate, FDR < 0.05) differential expression in TZ compared to SW, 1021 were
252
upregulated (Table S3) and 652 were downregulated (Table S4). Gene ontology
253
(GO) functional enrichment analysis (Table S5) indicated that GO categories 9 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
254
like biosynthetic and secondary metabolic processes and response to stimulus
255
were significantly over represented in the TZ. We focussed on pinosylvin
256
biosynthesis (secondary metabolism) and on its upstream pathways, then
257
extended our interest to other processes such as plant hormonal signalling,
258
dehydration, programmed cell death, transcription factors, cell wall modification
259
and defence.
260 261
Metabolic processes activated during heartwood formation
262
The biosynthesis of stilbenes is initiated by conversion of phenylalanine to
263
cinnamic acid, catalysed by the enzyme phenylalanine ammonia lyase (PAL).
264
STS converts the activated cinnamic acid (cinnamoyl-CoA) into pinosylvin. We
265
defined the TZ as the zone where STS encoding transcripts are most highly
266
expressed. Congruently, we could observe 5 to 13 fold upregulation of STS
267
encoding TCs in TZ compared to SW. Our transcriptome data also showed that
268
PAL encoding transcripts were 7 to 12 fold upregulated in the TZ. The next
269
enzyme, presumably an acyl-CoA ligase acting on cinnamate, has not been
270
characterized from pine, but some 4-coumarate:CoA ligase (4CL) transcripts
271
were upregulated with a 3 to 13 fold change. Lignin biosynthesis related
272
transcripts (including 4CL) were also upregulated in TZ. In the end of the
273
pathway, pinosylvin is in part methylated to PSME. Chiron et al. (2000a)
274
characterised in Scots pine an O-methyl transferase catalysing this reaction,
275
however, TC180420 that matches the published PMT sequence (99 % identity)
276
was not upregulated (in fact, very few reads mapped to it). Instead, another O-
277
methyltransferase encoding TC (TC166778) was upregulated with a 5 fold
278
change in TZ compared to SW (Figure 2, Table S6).
279 280
Upstream of stilbene pathway is the shikimate pathway that provides
281
phenylalanine for stilbene biosynthesis, as well as for the phenylpropanoid
282
pathway in general (including lignin and lignan biosynthesis) and for protein
283
synthesis. The shikimate pathway starts with the conversion of
284
phosphoenolpyruvate and erythrose 4-phosphate to 3-deoxy-D-arabino-
285
heptulosonate-7-phosphate (DAHP) by the enzyme DAHP synthase (DAHPS).
10 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
286
DAHPS is thought to be the rate-controlling enzyme of the shikimate pathway
287
(Tzin et al., 2012), and we observed 4 to 7 fold upregulation in TZ compared to
288
SW for TCs encoding DAHPS. Several other TCs encoding enzymes of the
11 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
289
shikimate pathway were upregulated in TZ, including those for 3-
290
dehydroquinate synthase (4 fold) and shikimate kinase (4-7 fold) and finally
291
chorismate mutase and prephenate dehydratase, 5 and 10-13 fold, respectively
292
(Figure 2, Table S6).
293 294
Further upstream, transcripts encoding phosphoenolpyruvate carboxykinase
295
was upregulated 5 to 8 fold in TZ. This enzyme converts oxaloacetate to
296
phosphoenolpyruvate in gluconeogenesis. TCs for enzymes involved in
297
glycolysis and the oxidative pentose phosphate pathway also showed
298
upregulation and finally transcripts for sucrose synthase were induced with a 2
299
to 5 fold change (Figure 2, Table S6). Activity of sucrose synthase is often taken
300
as indication of metabolic activity (Winter and Huber, 2000), showing that
301
metabolite production is taking place in the TZ before it is programmed for cell
302
death (see below). It further indicates that breakdown of sugar is then
303
channelled to the downstream processes like stilbene biosynthesis both for
304
building the carbon backbone and providing metabolic energy for the process.
305
Our transcriptome data show that stilbene biosynthesis definitely takes place in
306
situ in the TZ.
307 308
Plant hormones in the transition zone
309
The plant hormones ethylene and auxin have been suggested to be involved in
310
HW formation particularly their ratio being important (Hillis, 1968; Shain and
311
Hillis, 1973). Indeed, for TCs encoding of 1-aminocyclopropane-1-carboxylate
312
oxidase (ACO) we observed the highest differential expression ratio between
313
SW and TZ, an induction of 36-61 fold (Table S6). ACC synthase (ACS) was
314
not induced, but S-adenosylmethionine (SAM) synthase was expressed 3-9
315
times more strongly in TZ than in SW. Note that SAM synthase provides
316
substrate also for methylation of pinosylvin. Transcripts relating to auxin
317
biosynthesis or transport were not up or downregulated in TZ, but a group of
318
auxin response and auxin related or induced transcripts were downregulated
319
with an average of 2 fold change between TZ and SW (Table S6). Gibberellins
320
have not been suggested to be involved in HW formation. Beauchamp (2011)
12 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
321
was able to detect the inactive precursor GA24 in both SW and TZ, but not the
322
active forms or the 2-hydroxylated inactive forms. Still, one of the induced
323
transcripts encoding GA 2-oxidase (TC171916) was 16 times upregulated in TZ
324
compared to SW.
325 326
Dehydration and programmed cell death during heartwood formation
327
Moisture content in the TZ is low but still higher than in the HW. We observed
328
that a group of desiccation-related (DPR) transcripts was upregulated with fold
329
change of 23 to 31 in TZ (Table S6). DPR transcripts have been suggested to
330
be involved in tolerance of desiccation in plants (Zha et al., 2013; Piatkowski et
331
al., 1990). A group of aquaporin-like transcripts, on the other hand, were
332
downregulated (Table S6). Aquaporins are involved in regulation of cell-to-cell
333
water transport and in maintaining water homeostasis in plants (Bienert and
334
Chaumont, 2011; Hachez et al., 2006). Upregulation of the DPR and
335
downregulation of aquaporin transcripts may relate to the fact that water is
336
withdrawn from the TZ during HW formation. Transcripts encoding an S1-like
337
nuclease and a bifunctional nuclease (BFN) were highly induced in TZ with 13
338
and 20 fold change (Table S6) compared to SW. Both nucleases have been
339
described to be involved in PCD (Bollhöner et al., 2012; Farage-Barhom et al.,
340
2008). Upregulation of these transcripts indicated that PCD was taking place in
341
TZ.
342 343
Cell wall modification during heartwood formation
344
Cell wall modification and lignification related enzymes were upregulated during
345
HW formation. Lignification has been shown to be part of the HW formation
346
process in Scots pine (Bergström, 2003). Transcripts encoding enzymes
347
involved in lignification were upregulated in TZ at an average of 4 to 16 fold
348
(Figure 2, Table S6). Transcripts relating to cell wall carbohydrate modification
349
were expressed at an average 12 fold higher in TZ than in SW (Table S6).
350
13 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
351
Transcription factors in the transition zone
352
Two TCs encoding transcription factors, a MYB-like transcription factor and a
353
NAC domain-containing protein, were expressed 12 and 6 fold higher,
354
respectively, in TZ compared to SW (Table S6). Transcription factors of the
355
MYB family are often key regulators of secondary metabolite pathways (Dubos
356
et al., 2010). NAC domain proteins are involved in secondary cell wall
357
biosynthesis and PCD (Nakano et al., 2015; Bollhöner et al., 2012).
358 359
Resin acids and plant defence
360
Resin acids are conifer defence barriers to biotic challenges. They can be found
361
in the bark, in SW and in HW. According to Harju and colleagues (2003), HW
362
contains four times more resin acids on a volume basis than SW. It was
363
therefore surprising that transcripts involved in terpenoid biosynthesis and its
364
upstream pathways were neither up nor downregulated in the June
365
transcriptome data. In fact, very few reads mapped to these TCs (with an
366
average 95 % (and up to 99 %) identity with published diterpene synthase (Ro
367
and Bohlmann, 2006) and abietadienol/abietadienal oxidase (CYP720B)
368
(Geisler et al., 2016) sequences, indicating that constitutive synthesis of resin
369
acids was not taking place either (Table S7). On the other hand, a set of
370
pathogenesis-related transcripts were upregulated 5 to 20 fold (Table S6).
371 372
A set of unknown transcripts was also upregulated and downregulated in our
373
transcriptome data, sometimes up to 12 fold (Table S6).
374 375
Year round expression profiles
376
The Scots pine RNA-Seq data collected in June shed light to the metabolic
377
activities in the TZ during active growth of the tree. It did not answer the
378
question of when, during the yearly cycle, HW formation is most active, and left
379
completely open both the time and place of resin acid biosynthesis. Based on
380
the June results, we selected transcripts upregulated in TZ and apparently
381
related to HW formation as well as to resin acid biosynthesis for a year round
382
study using quantitative RT-PCR (qPCR) (Table S8).
14 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
383 384
We studied the expression profiles in the SW and TZ of increment cores of two
385
trees, harvested monthly in Punkaharju from March 2011 to March 2012. The
386
January 2012 increment cores were lost due to fragmentation of the frozen
387
tissue upon removal from the drill core.
388 389
PCR efficiency was determined for each primer pair used in the analysis, so the
390
qPCR results reflect relative transcript levels also between different genes.
391
Total RNA yields of the test trees varied, but not systematically throughout the
392
year (Table S9). Level of actin and histone transcripts compared to ribosomal
393
RNA (i.e., total RNA amount) in the extracted RNA samples varied remarkably
394
little (Table S10). Level of each selected transcript was expressed relative to the
395
geometric mean of the actin and histone transcripts and is shown throughout
396
the year in Figures 3 and 4.
397 398
We determined by qPCR expression levels of the selected transcripts also in
399
the SW and TZ RNA samples of the four trees from which the RNA-Seq data
400
was generated and differential expression was calculated, The two data sets
401
are in good agreement (Table S11).
402 403
Pinosylvin biosynthesis pathway transcripts showed a concerted
404
expression profile
405
The year round expression profiles of the two analysed trees varied, so they
406
were not averaged but are shown separately. Still, many processes important
407
for heartwood formation shared correlated patterns (Figures 3 and 4). The
408
expression of the stilbene biosynthesis pathway transcripts (PAL, 4CL, STS,
409
OMT) showed statistically significant correlation (Figure 5). Shikimate pathway
410
(DAHPS) followed the stilbene pathway closely, as did sucrose synthase
411
transcript levels. Transcript levels for all these markers in TZ increased in the
412
spring (Figure 3). Subsequently, they all dropped in June and showed recovery
413
during the summer months. PAL, 4CL, STS and OMT transcripts were
414
expressed at very low level in the SW throughout the summer months. The
15 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
415
expression level of these transcripts differed between the two trees in autumn
416
months and, interestingly, the SW followed the pattern although at a several fold
417
lower level. The level of all of these transcripts showed increase during the
16 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
418
winter months in both the TZ and the SW, even when the temperature was
419
constantly below the freezing point (Figure 4H).
420 421
Expression profiles of plant hormone related transcripts varied
422
Many studies have concluded that ethylene would play an important role in HW
423
formation. In the June 2011 RNA-Seq data we observed that ACO was the
424
number one differentially expressed transcript in TZ, compared to SW. The year
425
round qPCR data showed that ACO very strongly reacted to as yet undefined
426
stimuli throughout the sampling scenes, sometimes more strongly in SW and
427
very differently in the two trees sampled (Figure 4A). The observed differential
428
expression of ACO in the RNA-Seq samples seems not to be a typical pattern
429
during heartwood formation.
430 431
The expression level of the auxin responsive (AUX) transcript in the TZ was
432
generally very low, and ceased completely in the winter (Figure 4B). This
433
transcript was chosen since it was downregulated in TZ in the RNA-Seq data,
434
and it indeed showed lower expression in TZ throughout the year.
435
17 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
436
The two transcription factors shared a similar profile
437
Expression of the MYB-like and NAC domain-containing protein encoding
438
transcripts were higher in TZ in both trees throughout the year (Figure 4C and
439
D). In both trees their expression pattern correlated very clearly with the stilbene
440
synthase pathway transcripts as well as those for DAHPS and sucrose
441
synthase (Figure 5).
442 443
Programmed cell death
444
A very clear feature of the PCD related BFN transcript was that it was absent in
445
SW. In both trees, the transcript in TZ showed increased accumulation in May,
18 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
446
a trough in June and gradual increase through the late summer, in synchrony
447
with the stilbene pathway. BFN was not following the increased stilbene
448
biosynthesis peaks in late autumn and winter, suggesting that different signals
449
were inducing the stilbene pathway at that time (Figure 4E).
450 451
Expression profiles of resin acid biosynthesis genes
452
Transcripts involved in terpenoid biosynthesis were expressed at low level in
453
the June transcriptome data, both in SW and TZ. Excitingly, we noticed that the
454
transcript levels of diterpene synthase and abietadienol/abietadienal oxidase
455
(CYP720B) encoding transcripts (Figure 4F and G) in both SW and TZ started
456
to rise in early spring, dropped in late spring or early summer, and remained low
457
during the summer, autumn and winter. The expression level for these
458
transcripts was greater in the SW than in the TZ. The pattern was similar in both
459
trees.
460 461
Correlation between months of the year
462
We also calculated correlations between each sampling month for the
463
expression pattern of the selected genes (Figure S4). We expected that highest
464
correlation would be observed between successive months, and the TZ
465
samples of both trees followed this pattern roughly. On the contrary, the SW
466
samples showed strong correlations across the months, i.e. between months
467
that are far apart in the yearly cycle (Figure S4). Interestingly, this means that
468
the TZ follows seasonal changes for its gene expression patterns much more
469
strongly than the SW does.
470 471
19 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
472
Discussion
473 474
Scots pine (Pinus sylvestris L.) wood is extensively used as sawn timber
475
because of its strength, straightness, and visual appearance. Particularly the
476
HW of Scots pine is valued for its properties such as dimensional stability and
477
natural durability. Extractives of HW give, e.g., colour and low hygroscopicity to
478
HW timber and are key determinants of the passive defence, i.e., resistance
479
against degradation of wood by microorganisms (Nakada and Fukatsu, 2012;
480
Harju and Venäläinen, 2006; Leinonen et al., 2008). Sampling of Scots pine
481
transcriptomes in the SW and TZ to HW by RNA sequencing in June and qPCR
482
throughout the year allowed us to identify the place and time during the yearly
483
cycle when these important components are synthesised and HW is formed.
484 485
From carbon skeletons to the heartwood phenolics
486
During HW formation, reserves of the SW are consumed, in Scots pine mainly
487
starch and triglycerides (Magel, 2000; Bergström, 2003; Nakada and Fukatsu,
488
2012). Magel (2000) concluded that the most important contributor for sucrose
489
cleavage was sucrose synthase followed by invertase. Upregulation of
490
transcripts in TZ for both of these enzymes was observed in our data. In
491
addition, transcripts relating both to glycolysis and oxidative pentose phosphate
492
pathway were expressed at an average 2 fold more in TZ than SW, suggesting
493
that the breakdown of sugars is channelled downstream for synthesis of
494
precursors to the shikimate pathway, and finally to the pinosylvin pathway.
495 496
Phosphoenolpyruvate carboxykinase, an enzyme that converts oxaloacetate to
497
phosphoenolpyruvate (Bowsher et al., 2008), was found upregulated strongly in
498
TZ. In plants, oxaloacetate is generated from malate by malate dehydrogenase
499
in the cytosol. Malate is synthesised from breakdown of plant triglycerides via
500
the glyoxylate pathway in plant fat storage tissues (glyoxysomes). Study of
501
Saranpää and Höll (1989) and Bergstöm (2003) showed that amount of free
502
fatty acids was increased in the TZ and HW but not in SW. In addition, lipophilic
503
droplets were observed in cells of TZ associated with HW formation (Taylor et 20 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
504
al., 2002; Kampe and Magel, 2013). Upregulation of phosphoenolpyruvate
505
carboxykinase may indicate that triglycerides are used not only as a direct
506
source of heartwood extractives (free fatty acids), but also as energy and
507
carbon source for stilbenes.
508 509
During HW formation, the whole shikimate pathway is expressed at an average
510
of 2 fold higher in TZ and the rate limiting enzyme DAHPS up to 7 fold higher.
511
This pathway connects the primary metabolism (aromatic amino acid
512
biosynthesis) to the phenylpropanoid pathway, including the stilbene and lignin
513
biosynthesis pathways. Both of these branches are active in TZ, the stilbene
514
pathway being responsible for generation of the pine specific heartwood
515
extractives pinosylvin and its monomethyl ether. Pinosylvin lacks hydroxylation
516
of the phenylalanine derived aromatic ring, similar to cinnamic acid.
517
Furthermore, the Scots pine STS prefers cinnamoyl-CoA to 4-coumaroyl-CoA
518
(Fliegmann et al., 1992; Kodan et al., 2002). Only few plant 4CLs activate
519
cinnamic acid (Kumar and Ellis, 2003) and in some plant species activity of
520
enzymes with a more distant relationship to 4CL have been shown to be
521
involved in cinnamoyl-CoA formation (Gaid et al., 2012; Klempien et al., 2012).
522
We observed a possible 4CL-like candidate from our transcriptome data that
523
might be specific to pine pinosylvin biosynthesis during HW formation.
524
Nevertheless, characterization its enzymatic specificity is needed to address
525
this hypothesis.
526 527
The end product of pine stilbene biosynthesis is PSME, catalysed by PMT. A
528
Scots pine PMT was purified and partially characterised by Chiron et al.
529
(2000a). However, no transcript encoding this PMT was upregulated in our
530
transcriptome data, but instead another O-methyltransferase encoding
531
transcript was strongly induced in TZ. This finding suggests that the PMT
532
characterised by Chiron and colleagues might not be the correct Scots pine
533
PMT for pine HW extractive biosynthesis.
534 535
The strong upregulation of transcripts encoding enzymes leading to stilbene
21 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
536
biosynthesis in TZ indicates that these HW extractives are synthesised in situ in
537
the TZ, gaining the carbon skeleton from sucrose. In this respect, Scots pine
538
heartwood formation is of type I, or the Robinia-type (Magel, 2000).
539 540
Biosynthesis of heartwood resin acids
541
In addition to stilbenes and free fatty acids, Scots pine HW contains abundantly
542
resin acids (Harju et al., 2002; Fries et al., 2000). Contrary to expectations,
543
resin acid biosynthesis was neither up nor downregulated in TZ compared to
544
SW. Lorio (1986) proposed the concept of a balance between growth and
545
differentiation for resin acid biosynthesis in southern pines. He suggested that
546
during the growing season, the sink of photosynthates is preferentially the
547
growth process, and only after the growing period, resources are available for
548
resin acid production. The year round expression profile of transcripts encoding
549
diterpene synthase and abietadienol/abietadienal oxidase (CYP720B) showed
550
very low expression during summer months but a strong peak in early spring.
551
Throughout the year, resin acid biosynthesis related transcripts accumulated
552
more in SW than in TZ, still resin acids accumulate four times higher in HW than
553
in SW (Venäläinen et al., 2003). Resin acids do not seem to be biosynthesised
554
in situ in the TZ like stilbenes are, but instead they get loaded from SW to HW
555
during HW formation. Therefore, from the resin acid biosynthesis point of view,
556
Scots pine HW is of type II, or the Juglans type (Magel, 2000).
557 558
Balance between ethylene and auxin
559
The plant hormones ethylene and auxin have been suggested to be closely
560
related to HW formation. A recent study by Beauchamp (2011) reported that
561
auxin concentration was either reduced or below level of detection in the TZ, in
562
line with Hillis’s (1968) description that depletion of auxin content is associated
563
with HW formation. Shain and Hillis (1973) suggested that the balance between
564
ethylene and auxin during HW formation is important. Indeed, transcripts
565
encoding ACO were most strongly differentially expressed in TZ compared to
566
SW. On the other hand, transcripts for ACS, the rate limiting enzyme for
567
ethylene biosynthesis in most plants (Woeste et al., 1999; Hudgins et al., 2006),
22 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
568
was not induced. This may be a reflection of the fact that ACS regulation takes
569
predominantly place at protein level (Chae and Kieber, 2005), which we did not
570
address in our experiments. Transcripts for SAM synthase, the second
571
substrate for ACS (Yoon and Kieber, 2013), were induced in TZ, but SAM is
572
also needed for pinosylvin biosynthesis. The importance of ethylene in HW
573
formation, at least as a sufficient inducer, can however be questioned since
574
ACO was expressed erratically with little correlation between trees or zones.
575
Although we designed repeated sampling of the trees to avoid wound induction
576
(Figure S5), we cannot exclude that this caused a reaction in ACO. Transcripts
577
related to auxin biosynthesis were not differentially expressed between SW and
578
TZ. Again, regulation of auxin levels prominently take place after biosynthesis of
579
the hormone (Woodward and Bartel, 2005). Nevertheless, several auxin
580
response transcription factors were downregulated in TZ.
581 582
Transcriptional control in the transition zone
583
Two transcription factor encoding transcripts, for a NAC domain and a MYB-like
584
protein, were found to be significantly upregulated in the TZ and their
585
expression correlated with the stilbene pathway transcripts throughout the year
586
in both sampled trees. NAC domain transcription factors have been reported to
587
be involved in developmental processes, for example, in secondary cell wall
588
biosynthesis as well as biotic and abiotic stresses, acting as master switches to
589
regulate downstream genes (Nakano et al., 2015). The upregulated NAC
590
domain protein is however not similar to the white spruce (Picea glauca)
591
secondary cell wall biosynthesis NAC transcription factors, encoded by
592
PgNAC4 and PgNAC7 (Duval et al., 2014). Also the MYB-like transcription
593
factor is not particularly similar to the Pinus taeda secondary cell wall
594
biogenesis related MYB transcription factors that were isolated from
595
differentiating xylem, PtMYB1 (Patzlaff et al., 2003b) and PtMYB4 (Patzlaff et
596
al., 2003a). Still, as MYB transcription factors are in many cases main
597
regulators of secondary metabolism in plants, involvement of the upregulated
598
MYB transcript as well as the NAC transcript in stilbene or lignin biosynthesis
599
deserves a closer look.
23 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
600 601
Cells in the transition zone prepare for death
602
Pine heartwood is dead tissue and during HW formation the cells in the TZ not
603
only are responsible for filling the HW with extractives, but they also prepare for
604
their own death.
605 606
Moisture content in the TZ is lower than in SW (Hillis, 1987; Bergström, 2003).
607
The upregulation of transcripts encoding desiccation-related proteins in TZ
608
could result from the fact that water is withdrawn from the cells. The role of plant
609
DPR proteins may help the cells to tolerate the dehydration condition and allow
610
them to continue with the still necessary physiological processes during HW
611
formation. Downregulation of a group of aquaporin-like proteins may also relate
612
to dehydration in the TZ.
613 614
Transcripts encoding enzymes necessary for lignification, as well as groups of
615
cell wall polysaccharide modifying enzymes were expressed in the TZ during
616
HW formation. The cell walls of ray parenchyma cells in the TZ are indeed
617
undergoing lignification, as observed by Bergström (2003) and Yamamoto
618
(1982).
619 620
A chitinase and set of plant pathogenesis-related transcripts were upregulated
621
in the TZ. In general, these plant defence related proteins are induced in
622
response to biotic and abiotic stresses (Van Loon, 1997). Chitinases have been
623
reported to be involved not only in plant defence but also in developmental
624
processes, such as secondary cell wall lignification, and they are induced by
625
plant hormones (Grover, 2012). Yoshida et al. (2012) and Yang et al. (2004)
626
also reported that a set of pathogenesis-related proteins are highly expressed in
627
the TZ of Cryptomeria and Robinia, respectively.
628 629
The storage compounds in the TZ are depleted and consumed and the
630
parenchyma cell walls are undergoing modification. At the end of the process,
631
all physiological processes will eventually cease.
24 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
632 633
Programmed cell death is the key for heartwood formation
634
HW formation can been seen as a form of tissue senescence (Spicer, 2005;
635
Kampe and Magel, 2013). The BFN is an endonuclease similar to S1-like
636
nucleases (Farage-Barhom et al., 2008), is induced during senescence and is
637
involved in plant developmental PCD (Farage-Barhom et al., 2008; Bollhöner et
638
al., 2012). Farage-Barhom et al. (2008) demonstrated that the arabidopsis
639
BFN1 moves to the nucleus during leaf senescence and the protein was
640
detected from fragmented nuclei in the late stage of leaf senescence (Bollhöner
641
et al., 2012). Degradation of nuclei of ray parenchyma cells in the TZ has been
642
observed in past studies (Taylor et al., 2002; Spicer, 2005; Nakada and
643
Fukatsu, 2012). Thus, the upregulation of the BFN transcript in TZ, compared to
644
SW, is in line with the fact that PCD is involved in HW formation. Expression of
645
the PCD transcript BFN in TZ appears to be a key marker to distinguish
646
developmentally programmed HW formation in Scots pine from stress induced
647
extractive biosynthesis.
648 649
Timing of heartwood formation in Scots pine
650
The year round expression profiles of selected transcripts related to primary and
651
secondary metabolic pathways, the pinosylvin biosynthesis pathway, plant
652
hormones, transcription factors, programmed cell death and resin acid
653
biosynthesis pathway were investigated with the qPCR approach in order to
654
understand if there is a preferential time during the year when HW is formed.
655 656
As a general observation, we noticed that the expression levels in each tested
657
trees varied in the autumn and in the winter, but shared a pattern during the
658
summer months. It is apparent that the trees are responding both to shared
659
developmental programmes (e.g., HW formation), as well as to local challenges
660
that may be different for each tree individual responded to differently by different
661
genotypes. The two trees analysed were growing close to each other but were
662
not related.
663
25 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
664
The year round expression profiles of the two transcripts related to resin acid
665
biosynthesis were well in synchrony. The genes were expressed, and
666
presumably synthesis proceeded, in early spring and was practically shut down
667
for the summer months, in accordance to the concept suggested by Lorio
668
(1986). Resin acid biosynthesis transcripts were expressed in both SW and TZ,
669
but more strongly in SW, indicating that the compounds loaded in HW are
670
mainly not synthesised in situ in the TZ.
671 672
The ACO transcripts, strongly differentially expressed in TZ compared to SW
673
and related to the suggested ethylene involvement in HW formation, were
674
expressed at very high levels in both SW and TZ in one of the two trees. This
675
may reflect a challenge this tree met in the late summer. Usually, however,
676
regulation of ethylene synthesis is addressed to protein level control of ACS, not
677
ACO. The auxin responsive transcript was expressed in the spring and summer
678
in SW, with low levels in TZ throughout the year.
679 680
Transcripts responsible for stilbene biosynthesis, our signature pathway for HW
681
formation, were always more strongly present in TZ than in SW. There seems to
682
be a shared peak of activity in early summer, somewhat later than that for resin
683
acid biosynthesis. In late summer and in winter the two trees were behaving
684
nonsynchronously, but the pathway, including its upstream components of
685
sugar metabolism and aromatic acid biosynthesis, were reacting in concert.
686
Expression of the two transcription factor transcripts investigated strongly
687
correlated with expression of the stilbene pathway genes. An interesting
688
peculiarity is that transcripts for the stilbene pathway enzymes, and for the two
689
transcription factors, accumulated during the winter months even when the
690
maximum daily temperature kept well below zero for nearly two months (Figure
691
3 and 4H). This does not seem to be an artefact, other transcripts did not show
692
this and RNA yields from the tissues do not explain this phenomenon, nor does
693
the fraction of actin or histone transcripts compared to total RNA and used as
694
reference.
695
26 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
696
In a summary, we believe that we can answer the question when HW formation
697
takes place in Scots pine: Resin acids are formed early in the spring, followed
698
by stilbenes, whose synthesis is carried out through the summer. Programmed
699
cell death of TZ cells, i.e., their transformation into HW, takes place in autumn,
700
when growth has ceased. In November, everything is over and the outermost
701
dry zone of the tree trunk contains cells that prepare for the final active period in
702
their life, taking place the next summer.
703 704
Conclusion
705
Past studies have suggested that ethylene plays an important role in HW
706
formation. Upregulation of ACO in TZ seemed first to support this, but since
707
ACO was found to react strongly also in SW during the growth season, the
708
evidence is not conclusive.
709 710
The carbon skeleton for the HW stilbene biosynthesis is obtained by breaking
711
down sucrose and is channelled to pinosylvin biosynthesis through the
712
shikimate pathway. The result demonstrates that the Scots pine HW stilbenes
713
are synthesised in situ in the TZ.
714 715
Two transcription factors follow closely phenylpropanoid biosynthesis in the TZ,
716
leading not only to stilbenes but also to lignins. Finally, expression of a PCD
717
regulator distinguishes developmentally regulated heartwood stilbene
718
production from other activating signals (stilbenes are also stress metabolites).
719 720
Stilbene biosynthesis clearly takes place in TZ cells, but resin acids, another
721
prominent group of heartwood extractives are mainly synthesised in the
722
sapwood, based on steady state levels of transcripts encoding enzymes needed
723
for their biosynthesis. Therefore, heartwood formation is of type I from the
724
stilbene point of view, but type II from the resin acids point of view.
725 726
Our transcriptome data indicates that heartwood formation takes place during
27 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
727
the growth period, with different timing for resin acid and stilbene biosynthesis.
728 729
Materials and methods
730
Plant material for RNA Sequencing and qPCR studies
731
Increment cores for RNA-Seq were sampled in June 2011 from four 46 years
732
old Scots pine trees (T285, T328, T335 and T415) at Leppävirta progeny trial
733
stand number 30902 (62°25′N, 27°45′E), established by the Finnish Forest
734
Research Institute, METLA. On average, thirty increment cores (5 mm in
735
diameter) were randomly drilled through each tree trunk with a battery driven
736
increment core borer approximately at chest height level (130 cm) avoiding the
737
knots. The border between wet and dry wood was marked with a pencil upon
738
collection, after which the increment cores were frozen in dry ice and
739
transferred to the -80 °C laboratory freezer.
740 741
Increment cores for initial tests (including STS RT-PCR) throughout the summer
742
months June to August in 2010 were sampled from 35 years old Scots pine
743
trees at Punkaharju progeny trial stand number 62201 (61°49′N, 29°20′E) with
744
the aforementioned method (trees 11 and 20). Increment cores for the year
745
round expression study were sampled from other two 35-36 years old Scots
746
pine trees (trees C and D) at the same progeny trial in Punkaharju from March
747
2011 to March 2012. Samples were harvested in the middle of each month.
748
Four increment cores were drilled through the pith each month, about one cm
749
distance apart. During each sampling date, the sampling position was moved
750
laterally and longitudinally around the trunk (Figure S5). By this procedure we
751
minimised the effects of the previous drills on the samples. The sampling took
752
place from several interwhorls starting from around 180 cm downwards. The
753
increment cores were frozen in liquid nitrogen and stored at -80 °C.
754 755
Localisation of the transition zone
756
PS and PSME give strong fluorescence under UV excitation (Harju et al., 2009),
757
allowing straightforward visualisation of HW extractives. The frozen increment
758
cores were illuminated at wavelength 365 nm and photographed. The TZ
28 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
759
annual ring of each increment core was identified according both to the pencil
760
mark made in the field and the fluorescence image made in the laboratory. We
761
discarded increment cores where the pattern was not regular, for example when
762
stilbenes were present outside of the dry zone in SW, indicating a reaction to a
763
biotic or abiotic challenge.
764 765
Total RNA extraction
766
Annual rings from TZ and SW were excised with a scalpel while keeping the
767
increment core frozen. SW sample consisted of four to five pairs of annual rings
768
located at approximately four or five annual rings away from the TZ. Samples
769
were ground to fine powder in liquid nitrogen using first mortar and pestle and
770
subsequently a milling machine (Qscillating Mill MM400 Retsch GmbH,
771
Germany).
772 773
Total RNA isolation for SW and TZ samples was carried out as described by
774
Chang et al. (1993) with minor modifications. After overnight lithium chloride
775
precipitation, the RNA pellet was collected by centrifugation at 10 000 ×g for 30
776
minutes at 4 °C, washed with 70 % cold ethanol, air-dried for 3 minutes and
777
dissolved in 50 µL nuclease-free Milli-Q water. Genomic DNA was digested with
778
DNase I, followed by purification with RNeasy Plant Mini Kit (Qiagen, Germany)
779
according to the manufacturer’s protocol. The concentration and the quality of
780
the genomic DNA-free total RNA was assessed with spectrophotometer
781
(GeneQuant 1300 Eppendorf, Germany) and agarose gel electrophoresis.
782 783
Semiquantitative STS RT-PCR
784
Semiquantitative STS RT-PCR was performed for localisation of the TZ and
785
screening for an ideal of sampling time. The annual ring sampling and total RNA
786
isolation were performed as described above. First strand cDNA was
787
synthesised using 50 ng of total RNA with SuperScript III reverse transcriptase
788
(Invitrogen, USA) with oligo(dT)20 primers according to manufacturer’s
789
instructions in a 20 µL reaction volume. The STS RT-PCR was carried out with
790
the programme 94 °C for 2 min, 24 cycles of 94 °C for 30 s, 57 °C for 1 min and
29 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
791
72 °C for 1 min, and 72 °C for 7 min. Primers (Table S12) were designed to
792
span an exon-intron boundary with Primer3 software (Untergasser et al., 2007).
793 794
Sample preparation for RNA sequencing
795
Eight micrograms of genomic DNA free total RNA was selectively depleted from
796
ribosomal RNA (rRNA) according to the procedures described in RiboMinus
797
Plant Kit for RNA-Seq (Invitrogen, USA). 500 ng of rRNA-depleted RNA was
798
used for transcriptome library preparation. After the ligation of SOLiD adaptor
799
mix and reverse transcription, size fractionation of the transcriptome library was
800
performed with gel size selection of 150-200bp cDNA, followed by amplification
801
and purification according to the protocols described in SOLiD Total RNA-Seq
802
Kit (Invitrogen, USA). Next, paired-end sequencing was conducted with SOLiD
803
5500xl at Institute of Biotechnology, University of Helsinki.
804 805
Mapping of reads and data handling
806
The paired-end colour space reads of SW and TZ RNA-Seq libraries were
807
mapped against the Pinus EST collection, version 9.0 (The Gene Index
808
Databases, 2014) using the SHRiMP2 alignment tool (David et al., 2011) with
809
parameters as shown in Protocol S1. Results were obtained in the Sequence
810
Alignment/Map (SAM) format (Li et al., 2009) and then converted to sorted
811
Binary Alignment/Ma (BAM) format, duplicates were removed with Picard tools
812
(Wysoker et al., 2013). The abovementioned tasks were carried out on the
813
Finnish Grid Infrastructure clusters and supercluster servers in CSC-IT Center
814
for Science Ltd (Espoo, Finland).
815 816
Differential expression analysis
817
Tables of mapped read counts for SW and TZ RNA-Seq libraries were imported
818
to edgeR R session to build a DGEList object for differential analysis (Robinson
819
et al., 2010). Low expression transcripts of Pinus ESTs collection were filtered,
820
only transcripts that achieved at least eight counts per million (CPM) in at least
821
four libraries were kept. The trimmed mean of M-value normalization, dispersion
822
estimation and differential expression analysis was then carried out. List of the
30 Downloaded from www.plantphysiol.org on January 5, 2017 - Published by www.plantphysiol.org Copyright © 2016 American Society of Plant Biologists. All rights reserved.
823
statistically significant differentially expressed transcripts at false discovery rate
824
(FDR) of less than 0.05 were then exported to a text file for downstream
825
analysis.
826 827
The Pinus EST collection has annotations but as some of them are not
828
informative or are missing, we compared the sequences to databases using the
829
blastx algorithm (Altschul et al., 1997) and obtained annotation updates for
830
many cases. Original and updated annotations are shown in Tables S3 and S4.
831 832
Gene Ontology Enrichment analysis
833
Gene Ontology (GO) terms of Pinus EST collection was obtained using
834
Blast2GO (Conesa et al., 2005). GOSlim plant ontology mapping was then
835
carried out in order to reduce complexity of obtained GO annotations. GO term
836
enrichment analysis was performed for upregulated and downregulated
837
transcripts with GO annotated Pinus EST collection as referral using Fisher's
838
exact test with multiple testing correction of FDR function implemented in
839
Blast2GO. General enriched GO terms were filtered from result list at FDR