A fluorescence energy transfer-based mechanical stress sensor for specific proteins in situ

A fluorescence energy transfer-based mechanical stress sensor for specific proteins in situ
of 26
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
   A FRET based mechanical stress sensor for specific proteins insitu Fanjie Meng , Thomas M. Suchyna , and Frederick Sachs * Center for Single Molecule Biophysics, Department of Physiology and Biophysics, State Universityof New York at Buffalo  Abstract To measure mechanical stress in real time we designed a fluorescence energy transfer (FRET)cassette, noted stFRET, which could be inserted into structural protein hosts. The probe was madeof a green fluorescence protein (GFP) pair, Cerulean and Venus, linked with a stable α -helix. Wemeasured the FRET efficiency of the free cassette protein as function of the length of the linker, theangles of the fluorophores, temperature and urea denaturation and protease treatment. The linkinghelix was stable to 80C°, unfolded in 8M urea, and was rapidly digested by proteases, but in all casesthe fluorophores were unaffected. We modified the α -helix linker by adding and subtracting residuesto vary the angles and distance between the donor and acceptor, and assuming the cassette was arigid body we calculated its geometry. We tested the strain sensitivity of stFRET by linking bothends to a rubber sheet subjected to equibiaxial stretch. FRET decreased proportionally to the substratestrain. The naked cassette expressed well in human embryonic kidney (HEK-293) cells and surprisingly was concentrated in the nucleus. However, when the cassette was located into host proteins such alpha-actinin, non-erythrocyte spectrin and filamin A, the labeled hosts expressed welland distributed normally in cell lines such as 3T3 where they were stressed at the leading edge of migrating cells and relaxed at the trailing edge. When COL-19 was labeled near its middle withstFRET, it expressed well in C. elegans , distributing similarly to hosts labeled with a terminal GFPand the worms behaved normally. Keywords Förster resonance energy transfer; Relative orientation factor; Venus; Cerulean; Alpha-helix linker Mechanical stress is one of the most influential physical factors in biology and one of the leastcharacterized. While it is obvious from molecular dynamics [1–4] and force spectroscopy [5– 12] that forces deform molecules, the mechanics of cells is much more complicated, involvingthe interaction of heterogeneous polymers and membranes and their interaction with both twodimensional heterogeneous liquid membranes [13,14] and three dimensional cytoplasmicsolutions where signaling factors can vary in time and space [15–17]. Mechanical interactionsat the levels of cells, organs and organisms are responsible for such familiar physiology asmotor function, hearing [18], touch [19] and the regulation of blood pressure [20], but theinteractions are also deeply embedded in the biochemistry of the cell affecting such varied  process as the phenotype of stem cells [21], DNA transcription [22,23],, translation of cellular components by motor proteins such as kinesin [24], stress induced changes of structure asoccurs in shear stress modulation of the cytoskeleton of the endothelia [25,26] and more generalinteractions due to the physical chemistry of concentrated protein solutions[27]. To dissect * CorrespondenceFrederick Sachs. Center for Single Molecule Biophysics, Department of Physiology and Biophysics, State Universityof New York at Buffalo, Tel: 1-716-829-3289 extension 105, Fax: 1-716-829-2569, Email: sachs@buffalo.edu.  NIH Public Access Author Manuscript FEBS J  . Author manuscript; available in PMC 2008 June 1. Published in final edited form as: FEBS J  . 2008 June ; 275(12): 3072–3087. NI  H-P A A  u t  h  or M an u s  c r i   p t  NI  H-P A A  u t  h  or M an u s  c r i   p t  NI  H-P A A  u t  h  or M an u s  c r i   p t    which stresses affect which functions, we need labels that are sensitive to mechanical stressand that can be attached to specific proteins.To meet that need we designed a cassette (noted stFRET) that can be inserted into structural proteins and reports molecular strain via changes in FRET, and with appropriate calibration,molecular stress. The cassette consists of the GFP monomers Cerulean and Venus [28–32]linked by a stable α -helix [33]. This paper characterizes the properties of the probes, showsthat they can be efficiently incorporated into structural proteins such as collagen-19, non-erythrocyte spectrin, alpha actinin and filamin A within living cells, and that the FRET fromthis cassette changes with stress in situ .The efficiency of energy transfer for a FRET pair is E ∝ 1/(1 + (R/R  O ) 6 ) where R is the distance between the dipoles and R  O  is the characteristic distance for 50% energy transfer [34]. Themaximal sensitivity for changes in R occurs at R = R  O . For Venus and Cerulean R  O  is ~5nm[35] so we linked them with a 5nm α -helix. The efficiency is affected by the angle betweenthe transition dipoles as well as the distance between them, and estimated the probe geometry by varying the number of residues in the linker. Removing one residue caused a large changein angle with a small change in distance and adding or removing a full turn produced a changein distance with no change in angle. We used six mutants to solve for the three relevant anglesof the dipoles assuming the cassette was rigid. stFRET was stable over temperature and mildurea denaturing conditions, but with 8M urea the linker unfolded and the fluorophores remainedstable. Thus stFRET is robust.stFRET expressed well in various biological systems including 3T3 and HEK-293 cells and inC. Elegans. After insertion into a variety of structural host proteins such as collagen, filamin,actinin and spectrin it distributed in the same manner as the same hosts with terminal GFP tags.stFRET changed FRET with the spontaneous movement of motile cells, decreasing efficiencyin regions under tension and increasing in regions expected to be free of significant stress. Byaxially stretching C. Elegans we could demonstrate acute reversible changes in FRETassociated with tension and relaxation. stFRET opens the door to studying in real time many physiological processes that are modulated or driven by mechanical stress. Results General configuration and FRET spectra of stFRET and its variants Figure 2B shows the alignment of the DNA sequence of the linker with five modified versions(the predicted geometrical changes are in Table 1). As shown in Table 1, according to thegeneral property of alpha-helices one amino acid residue deletion produces a change ( − 100°)in angle with negligible change ( − 1.5 Å) in length. A five residue deletion of the helix rotatesthe structure 360° but shrinks the helix 2.7nm. Deletion or addition of two and half turns of the helix twists the structure − 180° or +180° and decreases or increases the length 1.35nm.Figure 2C gives the amino acid sequence and the segments of the helix linker we modified.Deletion of 18 amino acid residues eliminates five turns of the helix and a 9 amino acid deletioneliminates two and half turns. Figure 2A shows the general configuration of six stFRETvariants. Inward arrows show the excitation wavelength and outward arrows show the emissionwavelength. The width of arrows denotes light intensity.Figure 3A shows the emission spectrum of stFRET with excitation at 433nm. There are peaksat 475 and 527nm: with the 475 nm emission from the donor Cerulean and 527nm from theacceptor Venus with robust energy transfer. A 100 μ M solution of unlinked donor and acceptor (1:1 mixture, green filled squares and line with 433nm excitation) had a small emission at527nm due to the bleed-through from Cerulean, the donor (Blue filled reversed triangle andline) and some direct excitation of the acceptor Venus by 433nm (Black triangle and line). The Meng et al.Page 2 FEBS J  . Author manuscript; available in PMC 2008 June 1. NI  H-P A A  u t  h  or M an u s  c r i   p t  NI  H-P A A  u t  h  or M an u s  c r i   p t  NI  H-P A A  u t  h  or M an u s  c r i   p t    donor and acceptor mixture had E = 0 and D/Aratio = 2.47 ± 0.05 (Figure 3, B). However, for stFRET E = 44 ± 2.5% and D/Aratio = 0.47 ± 0.02, showing efficient energy transfer (for Eand D/ARatio calculation see Methods). Calibration of three angles and κ 2 Confident in the srcin of stFRET energy transfer, we purified the other five variants and measured their fluorescence (Figure 4A). All mutants exhibited robust FRET (Table 2). stFRETitself had 44 ± 2.5% energy transfer and the 5T construct had the highest efficiency E = 56 ±4.5%, the 2.5T construct increased E to 47 ± 2.1% while the 2.5I decreased E to 37 ± 0.9%.FT1AA and FT2AA, presumably only having their angles changed, decreased E to 29 ± 7.1%and 38 ± 4.3% (Figure 4B). Table 2 summarizes the apparent change of angles and distancesobtained by modifying the linker and the corresponding energy transfer efficiency.If we assume that a single residue alters the linker length by a translation of 0.15nm and 100degrees, and that the structure is rigid, we can use the data in Table 2 to solve for the probegeometry (see Experimental procedures). The numerical solutions gave θ A  = 3.83, θ D  = − 0.78and Φ  = 1.97 radians, and κ  2  = 0.86, which is 30% higher than 2/3, the κ  2  value one would obtain assuming of random rotation of the donor and acceptor (Figure 1). However, we wishto point out that a value of κ  2  ~2/3 does not necessarily imply the probes are moving randomly. Stability of the linker as perturbed by urea, temperature and proteinase K We did a number of tests to assess linker integrity. If the linker were an α -helix, then meltingwill increase the end-end spacing and the efficiency will decrease. With urea as a denaturant[36,37], Figure 5A shows that the efficiency of stFRET declined with concentration up to 8Mand the previously quenched donor emission recovered. Remarkably, the fluorophore spectrawere almost unaffected by urea with <10–15% change in amplitude (Figure 5 C and D). Figure5B shows that 1 to 8M urea caused the D/ARatio to increase from 0.46 to 1.21 as expected if the helix unfolded into a random coil allowing the donor and acceptor to move further apartand reducing energy transfer (Figure 5E).As a second test of the helix stability we tried to melt stFRET at elevated temperatures but the protein proved stable up to 80 C°. Figure 6A shows the temperature dependence of fluorescenceof 100 μ M stFRET protein excited at 433nm from room temperature to 80C°. Donor and acceptor emission both declined somewhat as the temperature increased probably due to adirect change in quantum efficiency but there was no significant change in transfer efficiencyfrom 60C° to 80C°, the upper limit of our measurements so that the linker structure can beconsidered quite robust.As a final test of linker integrity we digested stFRET with proteases that cut the linker but leftthe fluorophores intact. Figure 7 shows that proteinase K led to a rapid fall in efficiency thatwas complete within 1min. The D/ARatio changed from 0.42 to 1.95 over 30min (Figure 7B)compared a change from 0.46 to only 1.21 when the protein was treated with 8M urea (Figure5B). Similar behavior was found for all six constructs (data not shown). The donor and acceptor fluorophore spectra were unaffected by proteinase K after 30min digestion (Figure 7 C and D).Figure 5E and 7E are diagrammatic models summarizing the energy transfer between donor and acceptor under different treatments (the width of the arrows represents signal intensity). In Vitro  measurement of strain sensitivity To verify the strain responsiveness we bonded the ends of derivatized stFRET to a siliconerubber sheet using StreptagII-Streptactin™ and stretched the sheet equibiaxially on thefluorescence microscope. When the C- and N-terminal ends of stFRET were derivatized sothat it would be stretched with the sheet there was a reversible ~11% decrease in the D/A Ratio Meng et al.Page 3 FEBS J  . Author manuscript; available in PMC 2008 June 1. NI  H-P A A  u t  h  or M an u s  c r i   p t  NI  H-P A A  u t  h  or M an u s  c r i   p t  NI  H-P A A  u t  h  or M an u s  c r i   p t    (Figure 8). As a control we measured FRET from stFRET that was derivatized at one end onlyso that it was simply immobilized but not stretched and there was no significant change inFRET with strain (Figure 8). Non-specific binding of double tagged stFRET to an untreated silicone surfaces also produced no significant change in FRET with strain. Thus, stFRET issensitive to strain as expected from the solution assays and the design of the probe. Eukaryotic expression and targeting property of stFRET Before inserting stFRET into host proteins, we placed the gene under a eukaryotic promoter (Human cytomegalovirus) and transiently transfected HEK cells with stFRET alone. Controltransfections with Venus or Cerulean monomers showed no preferential localization and obviously no energy transfer (Figure 9 A, B, C, D, E, F). Cells transfected with stFRETdisplayed significant energy transfer (Figure 9I). stFRET localized to the nucleus with anextremely high density in the nucleoli (Figure 9K). Nuclear targeting proteins have a consensus amino acid sequence of Lysine/Arginine (K/R) 4–6  or smaller clusters separated by 10–12 amino acids: (K/R) 2 X 10–12 (K/R)3 [38]. The linker has multiple arginine clusters similar to nuclear targeting sequence, but simply removing oneor the other fluorophores from stFRET produced a uniform cytoplasmic distribution showingthat the linker’s sequence alone was not sufficient for targeting. These unexpected nuclear targeting properties of stFRET may provide a useful tool for understanding nuclear proteintransport. Host proteins of stFRET with normal expression showing stress sensitivity We have inserted stFRET into various host proteins including collagen-19 (Figure 10G), non-erythrocyte spectrin (Figure 10E), filamin A (Figure 10C), alpha-actinin (Figure 10A), and expression systems including HEK-293, 3T3, and C. Elegans,  and the insertion locations wereoptimized to obtain protein distributions similar to those observed for the host protein C-terminal tagged by GFP or Cerulean (Figure 10B, D, F and H). Inserting stFRET into host proteins eliminated nuclear targeting. The fluorescence of stFRET in cultured cells was located in the cytoplasm and/or the cell membrane depending upon the host (Figure 10A, C and E).We have expressed the construct of the most abundant collagen in C. Elegans , COL-19, and the protein was properly assembled showing the typical striated pattern and the worms behaved normally. When we stretched the worm with micromanipulators, the labeled COL-19 showed a decrease in FRET efficiency with stretch, and in convex regions as they actively wiggled (Figure 10G and H).Figure 11 indicates stFRET integrated into actinin and filamin can sense tension in situ .Migrating 3T3 cells have a characteristic leading and lagging edge and Figure 11A, B and Cshows the donor, acceptor and FRET images from three confocal microscopy channels. stFRETwas distributed evenly across the cytoplasm as visualized with a 16 color pseudo-color map(Figure 11C,D). Transfection with Actinin-stFRET revealed that during migration, the laggingedge showed higher energy transfer than the leading edge (Figure 11E and F), i.e., it wasrelaxed. We measured the efficiency of various domains in the lagging and leading edges fromfourteen confocal image stacks. The lagging edges (the red outline domain) nearly doubled theFRET efficiency compared to the leading edge (blue and green outline domains). Multiple cellshad the same behavior but because of the complexity of the various shapes it was difficult toarrive at any useful statistic for frequency. We have shown a typical cell with different domainsas an internal control. The same phenomena was observed in Filamin-stFRET transfected 3T3cells (Figure 11G, H, I, J, K and L). Figure 11G, H and I are three confocal image channels,and Figure 11J is the pseudo-color image of stFRET protein distribution. Figure 11K is theFRET efficiency image in which three domains were selected. The efficiency in the red outlined domain is twice as high as that in the blue and green domains (Figure 11L). These data suggest Meng et al.Page 4 FEBS J  . Author manuscript; available in PMC 2008 June 1. NI  H-P A A  u t  h  or M an u s  c r i   p t  NI  H-P A A  u t  h  or M an u s  c r i   p t  NI  H-P A A  u t  h  or M an u s  c r i   p t    tension in both actinin and filamin is lower in domains close to the lagging edge (whereadhesion to the substrate is released), and higher at the leading edge where adhesions pullingthe cell forward. Discussion Designed to be an in situ  stress sensor, stFRET has a robust and predictable energy transfer  both in vitro  and in vivo . We were able to explore the geometry of stFRET by perturbing thelinker length and terminal angles using the known properties of alpha helices. FRET efficiencychanged in a predictable manner with the postulated geometry suggesting that the fluorophoresare not free to rotate. A recent molecular dynamics simulation study of FRET in lysozymefound that κ   and R  O  could be correlated by as much as 0.8, so that FRET measurements thatassume random rotational freedom are likely to be in error [39]. The ability to change angleand distance by varying the linker can be used in vivo  to examine the effect of host proteins on probe geometry. Regardless of the coupling of the fluorophores to the linker, all of the host proteins we studied were coiled-coiled dimers or trimers so that the fluorophores of stFRETwould not be able to rotate freely.Figure 2 A shows the predicted mean structure of free stFRET. The three unknown angles of equation (6) were solved using data for the six mutants using the least squares equation solver in MAPLE. The solutions were stable to perturbations of the starting values suggesting thatwe were measuring a constrained system. Our final solution: θ A  = 3.83, θ D  = − 0.78, Φ  = 1.97yielding κ  2  =0.86. There will be bending and flexing motions of the structure in solution, butwe obtained consistent answers from the overdetermined set of equations suggesting that thecalculated mean values are at least self consistent. The geometric values we have calculated would represent mean values weighted by the efficiency. Fluctuations that bring the dipolescloser are more heavily weighted than those that move them further away although the probability of occupancy of these conformations is another weighting factor. A detailed MDsimulation would be useful, but that is not essential to the use of stFRET as a probe of molecular stress since the most important variables are the differences in efficiency, i.e. the gradients of stress.The robust nature of stFRET was clear from the melting experiments. stFRET was thermallystable up to at least 80°C with the FRET efficiency virtually unchanged. Melting the linker with urea (Figure 5) [40] left the fluorophores untouched (Figure 5 C, D), but decreased theenergy transfer consistent with unfolding of the linker (Figure 5 B, E). Two models have been proposed for urea-induced protein denaturation: the binding model, in which the denaturant binds weakly but specifically to sites exposed by the unfolded proteins [41], and a solventexchange model in which the interaction of the solvent and the denaturant is a one-for-onesubstitution reaction at particular sites [42]. stFRET might serve as a useful probe to examinethese alternatives.The sensitivity of stFRET to protease cleavage has both positive and negative implications. If  proteases are accidentally present in situ , they could cleave stFRET and provide misleadingresults. We saw no evidence of protease activity in HEK or 3T3cells or C. Elegans . However,the presence of intracellular proteases has been associated with acute pancreatitis, proposed toarise from trypsin over-activation in large endocytotic vacuoles of acinar cells [43]. Thus tostudy pancreatitis, stFRET may be a useful probe (cf., Figure 7).Figure 8. Double Streptag II tagged stFRET shows a decrease in D/A Ratio when stretched onsilicone rubber disks. stFRET immobilized with a single Streptag II tag did not respond tostrain, nor did double tagged stFRET bound to a non-derivatized rubber disk. The D/A Ratiowas monitored at 10 spots on each disk during equibiaxial strain and the results averaged. Only Meng et al.Page 5 FEBS J  . Author manuscript; available in PMC 2008 June 1. NI  H-P A A  u t  h  or M an u s  c r i   p t  NI  H-P A A  u t  h  or M an u s  c r i   p t  NI  H-P A A  u t  h  or M an u s  c r i   p t  
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks