Kaye B et al. (2017), Bridging length scales to measure polymer assem...

XB-ART-54124

Mol Biol Cell 2017 May 15;2810:1379-1388. doi: 10.1091/mbc.E16-05-0344.

Show Gene links Show Anatomy links

Bridging length scales to measure polymer assembly.

Kaye B , Yoo TY , Foster PJ , Yu CH , Needleman DJ .

???displayArticle.abstract???
Time-resolvable quantitative measurements of polymer concentration are very useful to elucidate protein polymerization pathways. There are numerous techniques to measure polymer concentrations in purified protein solutions, but few are applicable in vivo. Here we develop a methodology combining microscopy and spectroscopy to overcome the limitations of both approaches for measuring polymer concentration in cells and cell extracts. This technique is based on quantifying the relationship between microscopy and spectroscopy measurements at many locations. We apply this methodology to measure microtubule assembly in tissue culture cells and Xenopus egg extracts using two-photon microscopy with FLIM measurements of FRET. We find that the relationship between FRET and two-photon intensity quantitatively agrees with predictions. Furthermore, FRET and intensity measurements change as expected with changes in acquisition time, labeling ratios, and polymer concentration. Taken together, these results demonstrate that this approach can quantitatively measure microtubule assembly in complex environments. This methodology should be broadly useful for studying microtubule nucleation and assembly pathways of other polymers.

???displayArticle.pubmedLink??? 28356424
???displayArticle.pmcLink??? PMC5426851
???displayArticle.link??? Mol Biol Cell

???attribute.lit??? ???displayArticles.show???

	FIGURE 1:. The assumed conditions for FRET. Some tubulin molecules are labeled with a donor fluorophore (blue), some are labeled with an acceptor fluorophore (red), and some are unlabeled (gray). We assume that FRET occurs only between a donor-labeled tubulin molecule and a nearby acceptor-labeled tubulin molecule in a microtubule, whereas tubulin molecules not incorporated in microtubules are not engaged in FRET.
	FIGURE 2:. Measuring FRET with FLIM. (A) Schematic diagram of excitation and relaxation pathways of the donor (blue) fluorophore. When a donor fluorophore absorbs an incoming photon, the fluorophore is raised into an excited state. The fluorophore relaxes back to the ground state by either emitting a photon or releasing heat. FRET introduces an additional nonradiative pathway for the fluorophore to relax. Thus, the average time the fluorophore spends in an excited state, referred to as the lifetime, is shorter when the fluorophore is engaged in FRET. (B) In purified solutions, in the absence of acceptor, fluorescence lifetime is not significantly affected by polymerization. Histogram of photon arrival times from Atto565-conjugated tubulin (donor-labeled tubulin) in Taxol-assembled microtubules (red dots) and without Taxol (blue dots). Bayesian analysis of these histograms using a single-exponential decay model estimates the fluorescence lifetimes (with Taxol; 3.45 ± 0.04 ns; without Taxol, 3.67 ± 0.03 ns) and provides the corresponding models (with Taxol, red curve; without Taxol, blue curve). (C) In purified solutions, Taxol-induced microtubules formed in the absence of an acceptor (green) produce a photon arrival time histogram that is a decaying exponential with a lifetime of ∼4 ns. FRET happens in the presence of acceptor fluorophore (purple), which induces the addition of a short component with a ∼1-ns lifetime.
	FIGURE 3:. Investigating the relationship between FRET fraction and intensity. (A) An intensity image of microtubule structures in extract (left) and corresponding FRET-fraction map (right). (B) FRET fraction vs. intensity from the data in A for individual pixels (small blue dots) and grouped pixels (black dots). Error bars are the SD of the posterior distribution. The grouped pixels are well fitted by Eq. 7 (dark gray dashed line) with Pf = 0.123 ± 0.006 and , where error is the 95% confidence interval.
	FIGURE 4:. Fit parameters change as expected when acceptor concentration is varied. (A) Intensity images (left) and FRET-fraction maps (right) of Taxol-induced microtubules in Xenopus egg extracts with high (top) and low (bottom) acceptor concentrations. FRET-fraction maps were sensitive to acceptor concentration, whereas the intensity images showed no significant differences. (B) Colored dots, FRET fraction and intensity from the data in A for grouped pixels. Error bars are the SD of the posterior distribution. The grouped pixels are well fitted by Eq. 7 (gray dashed lines) with Pf = 0.107 ± 0.004 and for 1.3 μM acceptor, and Pf = 0.058 ± 0.004 and for 0.6 μM acceptor. Samples with more acceptor have a larger horizontal asymptote, leading to a larger Pf, the probability of FRET. Meanwhile, the x-intercept is unchanged, leading to , the number of photons from donor in monomer, being unchanged. (C) Black dots: Pf determined from model fitting as in B. Colored circles denote the Pf values from the best fit of data from samples in A and B. Error bars are 95% confidence intervals. Pf increases linearly with acceptor concentration (gray dashed line). (D) Black dots, , determined from model fitting as shown in B. Colored circles denote the values from the best fit of data from samples in A and B. Error bars are 95% confidence intervals. is unchanged when acceptor concentration is varied.
	FIGURE 5:. Fit parameters change as expected when acquisition time is varied. (A) Intensity images and FRET-fraction maps of Taxol-induced microtubules in Xenopus egg extracts acquired with a long (purple) and a short (green) acquisition time. Shorter acquisition times resulted in dimmer images but similar FRET-fraction maps. (B) Colored dots, FRET fraction and intensity from the data in A for grouped pixels. Error bars are the SD of the posterior distribution. The grouped pixels are well fitted by Eq. 7 (gray dashed line) with Pf = 0.330 ± 0.012 and = 76.4 ± 3.5 for 50-s acquisition, and Pf = 0.313 ± 0.031 and = 14.8 ± 1.9 for 10-s acquisition. The x-intercept increases with acquisition time, leading to a larger . Meanwhile, the horizontal asymptote, which determines Pf, is unchanged. (C) Black dots, Pf determined from model fitting as shown in B. Colored circles denote the Pf values from the best fit of data from samples shown in A and B. Error bars are 95% confidence intervals. Pf is unchanged when acquisition time is varied (gray dashed line). (D) Black dots, determined from model fitting as in B. Colored circles denote the values from the best fit of data from samples in A and B. Error bars are 95% confidence intervals. increases linearly with acquisition time (gray dashed line).
	FIGURE 6:. Microtubule dilution series to test measurements of polymer concentration. Polymer measurements by FRET-intensity (green dots) and centrifugation followed by absorption at 280 nm (blue dots) correspond to the expected polymer amount (black dashed line). Error bars are SEM. Inset, map of the concentration of tubulin subunits in polymer. Scale bar, 25 μm.
	FIGURE 7:. Investigating the relationship between FRET fraction and intensity in U2OS cells. (A) An intensity image of a mitotic spindle from a cell with both donor- and acceptor-labeled tubulin (top) and from a cell with only donor-labeled tubulin (bottom). (B) Colored dots, FRET fraction and intensity from the data in A for grouped pixels. Error bars are the SD of the posterior distribution. The grouped pixels from the sample with both donor and acceptor is well fitted by Eq. 7 (gray dashed lines) with Pf = 0.091 ± 0.008 and , where error is the 68% confidence interval.

References [+] :

Bastiaens, Fluorescence lifetime imaging microscopy: spatial resolution of biochemical processes in the cell. 1999, Pubmed