7
S1B) and CLSM (Fig 1A; bottom inset and Fig. S1C). Visualization of GFP-TUA6 labeled
139
microtubules is inherently compromised by background possibly owing to free tubulin dimers
140
containing GFP-TUA6. With SIM however such problems were significantly ameliorated and
141
GFP-TUA6 cortical microtubules were visualized with high contrast (Fig. 1B and Fig. S1D).
142
Furthermore, it was shown by SIM that GFP-TUA6 is incorporated to the microtubule lattice in a
143
discontinuous manner (Fig. 1B; top inset; Fig. S1D) which was not apparent in WF (Fig. 1B;
144
bottom inset; Fig. S1E). Such discontinuous incorporation of GFP-TUA6 tagged dimers in the
145
microtubule lattice (for mechanism see Waterman-Storer and Salmon, 1998) was best
146
appreciated with linear profiles drawn along GFP-TUA6 labeled microtubules. These showed
147
alternating, bell-shaped fluorescence intensities of variable amplitude, with clear peak-to-peak
148
separation by SIM (Fig. S2A inset and Fig. S2C) as compared to WF (Fig. S2B inset and Fig.
149
S2C). The read-out of such profiles (Fig. S2C) also showed that maximum fluorescence
150
intensities with SIM are well above background levels when compared to WF, further
151
substantiating the use of SIM in generating highly contrasted acquisitions in samples inherently
152
plagued with high background fluorescence as the GFP-TUA6 expressing hypocotyl cells. The
153
speckled appearance of GFP-TUA6-labeled microtubules probably reflects to the stochastic
154
incorporation of GFP-TUA6-containing dimers diluted in a soluble tubulin pool of unlabeled
155
dimers and it might prove a useful source for fluorescence speckled microscopy (Salmon &
156
Waterman, 2011).
157
Microtubule nucleation through lateral branching from walls of pre-existing microtubules, was
158
also followed by SIM, as it represents a major mechanism of microtubule nucleation in higher
159
plants (Chan et al., 2009). SIM allowed visualizing the very onset of branch formation and
160
release of nascent microtubules (Fig. 1C; top inset and Fig. S1F) while this was not possible with
161
WF (Fig. 1C; bottom inset and Fig. S1G).
162
In addition, SIM resolved in good detail the composite nature of cortical microtubule bundles,
163
especially in the mpk4 mutant (Fig. 1D; top inset and Fig. S1H), and outperformed WF (Fig. 1D;
164
middle inset and Fig. S1I) and CLSM (Fig. 1D; bottom inset and Fig. S1J) in this respect. The
165
discrimination of very proximal microtubules was best illustrated in orthogonal projections,
166
which provided a clear view on the resolution potential of SIM (Fig. 1E) as compared to WF
167
www.plantphysiol.org
on July 21, 2018 - Published by
Downloaded from
Copyright © 2014 American Society of Plant Biologists. All rights reserved.
8
(Fig. 1F) and CLSM (Fig. 1G) making possible to clearly discriminate such adjacent
168
microtubules as individual fluorescent spots with SIM (Fig. 1E) compared to WF (Fig. 1F) and
169
CLSM (Fig. 1G) which showed only one area of diffuse fluorescence.
170
The quantitative potential of SIM superresolution capacities was measured with the full width at
171
half maximum (FWHM) of normalized, fluorescence intensity profiles drawn perpendicular to
172
individual microtubules. Averaged FWHM values of GFP-MBD and GFP-TUA6 were then
173
compared to respective values obtained with WF and CLSM. During profiling of individual
174
microtubules, care was taken to use identical and co-aligned profiles, meaning that the maximum
175
intensity of all profiles was located at the same position (see Materials and Methods). Since the
176
software used (see Materials and Methods) gives different values for SIM, WF and CLSM, all
177
intensity profiles were normalized to values between 0-1. From the bell curves obtained in this
178
way, the FWHM corresponds to the width of the curve at a normalized intensity value of 0.5.
179
Taking into account the above, SIM outperformed WF and CLSM resolution. The FWHM for
180
GFP-MBD microtubules was 106±20 nm (mean
±SD; n=27; Fig. 1H) while for GFP-TUA6
181
microtubules it was 118±25 nm (mean
±SD; n=42; Fig. 1I). The difference between the mean
182
FWHM for GFP-MBD labeled microtubules and for GFP-TUA6-labaled microtubules is not
183
significant (p=0.141). The respective FWHM value for GFP-MBD microtubules visualized by
184
WF was 231±26 nm (mean
±SD; n=27; Fig. 1H) and for CLSM it was 235±20 nm (mean±SD;
185
n=27; Fig. 1J), while TUA6-GFP labeled microtubules were resolved at 223±44 nm (mean
±SD;
186
n=42; Fig. 1I) by WF and at 325±42 nm (mean
±SD; n=47; Fig. 1K) by CLSM.
187
For some acquisitions a 63
×1.40NA oil immersion objective was used to acquire SIM (Fig. 1C,
188
Fig. S3A, S3B) since similar objectives with similar numerical apertures are routinely used for in
189
vivo time-lapsed imaging of plant cells (e.g., Shaw et al., 2003 and van Damme et al., 2004 using
190
63
×1.20NA objectives; Vos et al., 2004 using 63×1.40NA). Although this objective allowed
191
subdiffraction acquisitions with SIM (with an average FWHM of 146
±24 nm (mean±SD; n=24;
192
Fig. S3B, S3C) this value was inferior to the 100
×1.57NA oil immersion objective which was
193
used exclusively thereon.
194
Quantitative analysis of cortical microtubule bundles
195
www.plantphysiol.org
on July 21, 2018 - Published by
Downloaded from
Copyright © 2014 American Society of Plant Biologists. All rights reserved.