11
in an effective time interval of 2.6 sec (detailed description on SIM adjustments for time lapsed
253
imaging are in Materials and Methods section). This interval is well within accepted published
254
range required to resolve microtubule dynamics (e.g., Shaw et al., 2003; Buschmann et al., 2010)
255
and it was uniformely used as a standard time interval to acquire time lapsed series of cortical
256
microtubule dynamics in all other microscopies used hereby (WF, CLSM, TIRF and SD) as
257
specified in Materials and Methods section. In all cases of imaging the acquisition settings were
258
adjusted towards optimal lateral resolution for the given time frame.
259
Again the spatial resolution was quantitatively defined by recording the FWHM of a number of
260
averaged, co-aligned and finally normalized transverse intensity profiles centered on individual
261
microtubules. For GFP-MBD labeled microtubules the recorded resolution of SIM was 135
±11
262
nm (mean
±SD; n=41; Fig. 4A), while for WF was 226±8 nm (mean±SD; n=41; Fig. 4C), for
263
CLSM was 238
±11 nm (mean±SD; n=41; Fig. 4E), for TIRF was 274±14 nm (mean±SD; n=41;
264
Fig. 4G) and for SD it was 323
±21 nm (mean±SD; n=54; Fig. 4I). With the SIM module GFP-
265
TUA6 labeled microtubules were resolved at 133
±8 nm (mean±SD; n=71; Fig. 4B) while the
266
respective resolution for WF was 225
±16 nm (mean±SD; n=71; Fig. 4D), for CLSM was 305±19
267
nm (mean
±SD; n=43; Fig. 4F), for TIRF was 283±19 nm (mean±SD; n=47; Fig. 4H) and for SD
268
was 309
±21 nm (mean±SD; n=83; Fig. 4J). The above values correspond to previously published
269
information on the resolution in above mentioned microscopic techniques (e.g., Zucker et al.,
270
1999; Wang et al., 2005). From these data it was concluded that the resolution of SIM, even after
271
the compromises for time-lapsed imaging, significantly exceeded that of WF, CLSM, TIRF and
272
SD. Since SIM clearly outperformed all other techniques, next sections are focused on time-
273
lapsed SIM for quantitative dynamic imaging of cortical microtubules.
274
In quantitative terms, cortical microtubules of hypocotyl epidermal cells stably transformed with
275
GFP-MBD or GFP-TUA6, exhibited alternating periods of growth succeeded by fast shrinkage
276
(e.g., Figs. 5A, B, E, F; Figs. S6A to S6D and Videos S1 and S2). This behavior of cortical
277
microtubules was best illustrated by respective kymographs (Figs. 5C, D, G, H). Moreover,
278
kymographs of GFP-TUA6 labeled microtubules exhibiting dynamics at both ends, revealed the
279
appearance of successive bright and dark stripes which remained vertical throughout the entire
280
observation time (Figs. 5G, H). Such stripes correspond to the discontinuous incorporation of
281
www.plantphysiol.org
on July 21, 2018
- Published by
Downloaded from
Copyright © 2014 American Society of Plant Biologists. All rights reserved.
12
GFP-TUA6 in the microtubule lattice as previously described (Figs. 1B, S1D, S2A and S2C; see
282
also Kner et al., 2009). The rates of growth and shrinkage were extrapolated from kymographic
283
analyses of the highly dynamic plus end and the less dynamic minus end. The respective growth
284
and shrinkage rate of the plus end of individual GFP-MBD-labeled microtubules were 6.15
±3.06
285
μm/min (mean±SD; n=65; Fig. 5O; Tables S1 and S2) and 17.65±7.39 μm/min (mean±SD;
286
n=65; Fig. 5O; Tables S1 and S2), while at the minus end GFP-MBD-labeled microtubules were
287
growing and shrinking at 1
±0.98
μm/min (mean±SD; n=41; Fig. 5O; Tables S1 and S2) and
288
0.97
±0.95
μm/min (mean±SD; n=33; Fig. 5O; Tables S1 and S2). In the case of GFP-TUA6-
289
labeled microtubules, the plus end was growing and shrinking at 7.84
±3.59
μm/min (mean±SD;
290
n=63; Fig. 5O; Tables S1 and S2) and 18.22
±5.74
μm/min (mean±SD; n=41; Fig. 5O; Tables S1
291
and S2) respectively. At the minus end the respective rates were 1.7
±1.74
μm/min (mean±SD;
292
n=23; Fig. 5O; Tables S1 and S2) and 1.19
±1.12
μm/min (mean±SD; n=26; Fig. 5O; Tables S1
293
and S2). As illustrated (Fig. 5O) and summarized in Tables S1 and S2, the above
measurements
294
are within previously published rates of microtubule growth and shrinkage for both constructs
295
(e.g., Dhonukshe and Gadella, 2003, Shaw et al., 2003, van Damme et al., 2004, Vos et al., 2004)
296
suggesting that SIM can provide a new tool for time lapsed imaging
of cortical microtubule
297
dynamics offering significantly higher resolution than commonly used techniques.
298
Frequently in kymographs from SIM recordings, short length growth and shrinkage events were
299
observed. Such length changes were ca. 200 nm (e.g., Fig. 5C) and they were smaller than the
300
resolution limits reported for
time-lapsed WF, CLSM, TIRF and SD. Such events were also
301
taken together for calculating catastrophe and rescue frequencies of individual microtubules
302
according to published procedures (Dhonukshe and Gadella, 2003). Moreover, catastrophe and
303
rescue frequencies were comparatively measured between SIM and WF because such images
304
were acquired simultaneously.
305
The overall catastrophe frequency of GFP-MBD-labeled microtubule plus ends was 0.020
306
events/s and the rescue frequency was 0.022 events/s when measured on WF acquisitions. With
307
SIM the catastrophe and rescue frequencies which were measured for the same microtubules as
308
for WF were 0.031 events/s and 0.033 events/s respectively (n=30 microtubules representing 47
309
minutes of observation).
310
www.plantphysiol.org
on July 21, 2018 - Published by
Downloaded from
Copyright © 2014 American Society of Plant Biologists. All rights reserved.