5
insight on the function of MAP65 proteins (Tulin et al., 2012; Portran et al., 2013; Stoppin-
83
Mellet et al., 2013), kinesin motors (Song et al., 1997), katanin-mediated microtubule severing
84
(Stoppin-Mellet et al., 1997) and microtubule dynamics (Moore et al., 1997). However, it is
85
explicitly acknowledged that such in vitro assays should be addressed in biologically coherent
86
systems with physiological relevance of microtubule dynamics (e.g., see Zanic et al., 2013 and
87
discussion in Gardner et al., 2013). Thus, an ideal approach would be to address microtubule
88
dynamics in the complex cellular environment at spatiotemporal resolutions achieved by in vitro
89
assays.
90
Subdiffraction optical microscopy techniques allow subcellular observations below Abbe’s
91
resolution threshold (Verdaasdonk et al., 2014), circumventing the need for TEM. Such
92
approaches permit dynamic subcellular tracking of appropriately tagged structures within living
93
cells (Tiwari & Nagai, 2013). Practically two superresolution strategies exist. The first involves
94
patterned light illumination, allowing superresolution acquisitions by two fundamentally
95
different methods, i.e., stimulated emission-depletion (STED; Hell, 2007) and structured
96
illumination microscopy (SIM; Gustaffson, 2000). The second interrogates the precision of
97
fluorophore localization and includes stochastic optical reconstruction microscopy (STORM;
98
Kamiyama & Huang, 2012) and photoactivation localization microscopy (PALM; Sengupta et
99
al., 2012). The above regimes differ in translational and axial resolution and their temporal
100
efficiency depends on the size of the imaged area. SIM is probably the best compromise for
101
superresolution live imaging as it offers reasonable lateral resolution (ca. 100 nm; Gustaffson,
102
2000) which may be reduced to 50 nm (Rego et al., 2012) and sufficient depth of imaging
103
combined with a reasonable axial resolution (ca. 200 nm). SIM allows dynamic imaging in a
104
broader field of view than STED, at biologically meaningful rates compared to PALM and
105
STORM (Kner et al., 2009) and with deeper imaging capacity compared to other superresolution
106
regimes and to TIRFM/ VAEM (Leung & Chou, 2011). Superresolution approaches have
107
received limited attention in plant cell biology field (Fitzgibbon et al., 2010, Kleine-Vehn et al.,
108
2011) and their resolution potential during live imaging was not quantified so far.
109
Hereby, high numerical aperture (NA) objectives are combined with SIM for acquisition and
110
systematic quantification of subdiffraction details of GFP-MBD (Marc et al., 1998; GFP fused to
111
www.plantphysiol.org
on July 21, 2018
- Published by
Downloaded from
Copyright © 2014 American Society of Plant Biologists. All rights reserved.
6
the microtubule binding domain of mammalian MAP4) and GFP-TUA6-labeled (Shaw et al.,
112
2003) cortical microtubules. For such studies, wild type plants and a mitogen activated protein
113
kinase 4 (mpk4) mutant, exhibiting extensive microtubule bundling due to the overexpression
114
and underphosphorylation of MAP65-1 (Beck et al., 2010) were used.
115
Results
116
General remarks on sample preparation
117
To optimize SIM
imaging of Arabidopsis thaliana hypocotyl epidermal cells, seedlings were
118
grown in darkness since etiolation promotes the thinning of the outer epidermal wall and reduces
119
the thickness of the cuticular surface (Kutschera, 2008). Thus hypocotyl surfaces can be wetted
120
well with acqueous mounting media used for
in vivo imaging (1/2 MS medium was used here),
121
excluding air pockets that may introduce additional refractive index mismatches upon imaging.
122
Meanwhile the contour of the etiolated hypocotyl surface is smoother allowing the wider
123
exposure of hypocotyl epidermal cell surfaces underneath the coverslip,
as required for high
124
numerical aperture objective with small working distance used hereby. Refractive index
125
mismatches pertaining to fluctuations in coverslip thickness were alleviated by using high
126
performance and low thickness tolerance coverslips. In addition, objective with high numerical
127
aperture (NA 1.57) was used in combination with immersion oil of high refractive index (1.66)
128
in order to obtain maximal
possible resolution during SIM, WF, CLSM, TIRF and SD imaging.
129
130
Microtubule organization
131
Since details of microtubule organization have been reported using other microscopies, the first
132
task was to characterize qualitatively and quantitatively the resolution potential of SIM and
133
provide relevant comparisons with WF and CLSM with respect to structural features of
134
microtubular arrays. This was particularly important as relevant literature is largely devoid of
135
quantitative resolution data using the above microscopies.
136
In this respect SIM revealed loop-like defects in GFP-MBD cortical microtubule bundles (Fig.
137
1A; top inset and Fig. S1A) which were indiscernible by WF (Fig 1A; middle inset and Fig.
138
www.plantphysiol.org
on July 21, 2018 - Published by
Downloaded from
Copyright © 2014 American Society of Plant Biologists. All rights reserved.