Enhancing the biocompatibility of rhodamine fluorescent
Download Enhancing the biocompatibility of rhodamine fluorescent
Preview text
Showcasing research from Dr. Gražvydas Lukinavičius laboratory of Chromatin Labelling and Imaging from Max Planck Institute for Biophysical Chemistry, Göttingen, Germany.
Enhancing the biocompatibility of rhodamine fluorescent probes by a neighbouring group effect
Our study demonstrates that the neighbouring group effect which results from positioning an amide group next to a carboxyl group in the benzene ring of rhodamines dramatically increases cell permeability of the rhodaminebased probes through stabilizing a fluorophore in a hydrophobic spirolactone state. Based on this principle, we create probes targeting tubulin, actin and DNA. Their superb staining efficiency allows long-term confocal and 3D Airyscan microscopy, and stimulated emission depletion (STED) nanoscopy with sub-30 nm resolution in living cells. Artwork prepared by Hartmut Sebesse and Gražvydas Lukinavičius.
As featured in:
See Gražvydas Lukinavičius et al., Chem. Sci., 2020, 11, 7313.
rsc.li/chemical-science
Registered charity number: 207890
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Chemical Science
EDGE ARTICLE
View Article Online
View Journal | View Issue
Cite this: Chem. Sci., 2020, 11, 7313 All publication charges for this article
have been paid for by the Royal Society of Chemistry
Received 15th April 2020 Accepted 20th June 2020 DOI: 10.1039/d0sc02154g rsc.li/chemical-science
Enhancing the biocompatibility of rhodamine fluorescent probes by a neighbouring group effect†
Jonas Bucevicˇius, a Georgij Kostiuk, and Graˇzvydas Lukinavicˇius *a
a Ru¯ta Gerasimaite˙,
a Tanja Gilatb
Fluorescence microscopy is an essential tool for understanding dynamic processes in living cells and organisms. However, many fluorescent probes for labelling cellular structures suffer from unspecific interactions and low cell permeability. Herein, we demonstrate that the neighbouring group effect which results from positioning an amide group next to a carboxyl group in the benzene ring of rhodamines dramatically increases cell permeability of the rhodamine-based probes through stabilizing a fluorophore in a hydrophobic spirolactone state. Based on this principle, we create probes targeting tubulin, actin and DNA. Their superb staining intensity, tuned toxicity and specificity allows long-term 3D confocal and STED nanoscopy with sub-30 nm resolution. Due to their unrestricted cell permeability and efficient accumulation on the target, the new probes produce high contrast images at low nanomolar concentrations. Superior performance is exemplified by resolving the real microtubule diameter of 23 nm and selective staining of the centrosome inside living cells for the first time.
Introduction
Modern super-resolution uorescence microscopy and nanoscopy techniques permit observation of biological processes in living organisms down to the molecular level.1 However, their success depends on the availability of biocompatible uorescent probes for specic labelling of cellular structures. Usually, these probes are constructed by coupling a uorescent dye via a linker to a small molecule ligand targeting a protein of interest. The linker and the uorophore make signicant contributions to the nal properties of such probes, oen resulting in compromised cell permeability and off-targeting.2 Recent studies dedicate a lot of attention to the rhodamine class uorophores that can cycle between a non-uorescent spirolactone and a uorescent zwitterion form. This property is exploited to generate uorogenic probes, which are sensitive to a number of environmental factors: pH, ion concentration, enzymatic activity, local microenvironment polarity or light.3 Since the spirolactone is a more hydrophobic molecule compared to the zwitterion, it makes rhodamine-based uorescent probes more cell-permeable.4 Regular rhodamines and carbon-substituted analogues (carbopyronines) have the
aChromatin Labeling and Imaging Group, Department of NanoBiophotonics, Max Planck Institute for Biophysical Chemistry, Am Fassberg 11, 37077 G¨ottingen, Germany. E-mail: [email protected] bDepartment of NanoBiophotonics, Max Planck Institute for Biophysical Chemistry, Am Fassberg 11, 37077 G¨ottingen, Germany † Electronic supplementary information (ESI) available. See DOI: 10.1039/d0sc02154g
equilibrium shied towards the zwitterion form, which results in relatively poor membrane permeability.4 To overcome this issue, several studies attempted to induce spirolactone form preference by introducing electron-withdrawing groups into the xanthene core5 or in the benzoic acid substituent.6 However, these approaches result in a bulkier core structure and alter the physicochemical properties of the dyes. In addition, all of them must rely on the uorogenicity in order to minimize background staining arising from the high working concentrations of the probes. We propose a way for modulating the spirolactone–zwitterion equilibrium without any change in the uorescent dye core structure by exploiting the neighbouring group effect (NGE), i.e. the phenomenon that two neighbouring carboxylic groups can inuence each other via steric, electrostatic or H-bond interactions.7 The NGE should not be confused with the term “neighbouring group participation” (NGP) – the direct interaction with a reactive center during a chemical reaction.8
Here, we introduce a new positional isomer with NGE to increase the cell permeability and uorogenicity of regular rhodamines without changing their spectroscopic properties. We demonstrate that excellent quality of wash-free multi-colour live-cell images can be obtained at low working concentrations with wide-eld, confocal and super-resolution microscopes even with moderately uorogenic probes. Selective centrosome staining is demonstrated using exceptionally well performing tubulin probes. Superior performance is exemplied by resolving the real microtubule diameter of 23 nm in STED nanoscopy images.
This journal is © The Royal Society of Chemistry 2020
Chem. Sci., 2020, 11, 7313–7323 | 7313
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Chemical Science
View Article Online
Edge Article
Results and discussion
Chemical synthesis of 40-carboxyl rhodamines
For a long time, the existence of a novel isomeric class of 40carboxyrhodamines was debated due to the synthesis challenge arising from steric hindrance and the ortho-effect altered reactivity of adjacent carbonyl groups.9 Isomeric mixtures of 50/ 60-carboxyrhodamines and uoresceins are usually obtained by the classical condensation process of phthalic anhydrides and 3-dimethylaminophenol or resorcinol respectively.10
The condensation between resorcinol and 1,2,3-benzenetricarboxylic acid or its homologs yields solely 70-carboxyuorescein and not 40-carboxyuorescein.11 Another widely applied and regioselective synthesis strategy, which can be applied in the synthesis of 40-carboxyrhodamines, involves addition of aryl-lithium species to the 3,6-substitutedxanthone.12 We used formation of di-tert-butyl ester 1 for the protection of 3-bromophthalic acid carboxyl groups for the halogen–lithium exchange reaction (Fig. 1). However, the tertbutyl protecting group has low tolerance to strong nucleophiles and the lithium–halogen exchange must be carried out at
temperatures between À90 and À120 C. Such temperatures are
not compatible with the reactivity of xanthones and their C/Si
analogues possessing strong electron donating dialkyl amino
groups at positions 3 and 6. When the halogen exchange was attempted at À78 C, mainly homo-addition products were
formed. To solve this issue, we performed the reaction at $À116 C with the more reactive silyl protected 3,6-dihydrox-
yxanthone 2 and its C (3) and Si (4) analogues and performed the late stage uorescein to rhodamine conversion strategy.13 To achieve this, TBDMS protecting groups on compounds 5–7 were removed with TBAF in THF and the obtained uorescein class compounds 8–10 were reacted with triic anhydride in the presence of pyridine in DCM to obtain the triates 11–13. The triate groups were converted to the desired (di)methylamino groups (compounds 14–16) by a Pd-catalyzed C–N crosscoupling reaction and the desired 40-carboxyrhodamines were
obtained by quantitative deprotection with a 1 : 1 mixture of DCM/TFA (Fig. 1). The 40-carboxyrhodamines behave differently
than other regioisomers upon activation with common peptide coupling reagents – the formed carbonyl activated species (NHS
ester, HOBt ester or similar) immediately undergoes
Fig. 1 Synthesis scheme of 4-TMR-COOH, 4-580CP-COOH, 4-610CP-COOH and 4-SiR-COOH dyes and conjugates 22–42. Their photophysical properties are depicted in the colour spectrum chart and the neighbouring group effect is highlighted.
7314 | Chem. Sci., 2020, 11, 7313–7323
This journal is © The Royal Society of Chemistry 2020
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Edge Article
View Article Online
Chemical Science
intramolecular cyclisation forming the anhydride (Fig. 1 and ESI Fig. 1†). In the presence of amines as nucleophiles the major products form via nucleophilic attack on a less hindered carbonyl group (position 40), although traces of spiroamide products can be observed by LC/MS analysis.
In vitro characterization of the neighbouring group effect in 40-carboxyrhodamines
40-, 50- and 60-Carboxyrhodamine regioisomers have almost identical absorption, uorescence emission spectra and quantum yield (QY) in phosphate buffered saline (PBS) (ESI Table 1†). In agreement with previously published observations, we obtained an excellent correlation between QY and uorescence lifetime (ESI Fig. 2a†).14 Even better correlation was obtained between uorophore brightness (product of QY and extinction coefficient) and uorescence lifetime with the exception of the 4-SiR-COOH dye, which showed a signicantly lower extinction coefficient (ESI Fig. 2b†).
To characterize the ability of the uorophores to switch between spirolactone and zwitterion states, we measured the absorbance in water–1,4-dioxane mixtures with known dielectric constants.15 We tted the absorbance change to the dose– response equation and obtained the dyeD50 value, which indicates the dielectric constant at which absorbance is halved (ESI Table 2 and Fig. 3†). dyeD50 of 4-SiR-COOH was signicantly
higher than the dyeD50 of 50- and 60-isomers. For all other uorophores, the dyeD50 of all three isomers were very close.
We have previously demonstrated that the performance of
the tubulin probes depends on the chosen taxane ligand, and that the optimal ligand is different for each uorophore.4a
Based on this knowledge, we synthesized a new series of tubulin probes by coupling the 40-, 50- and 60-regioisomers of TMRCOOH and 580CP-COOH uorophores to a Larotaxel derivative
(LTX-C8-NH2 (SI-4)) and 610CP-COOH and SiR-COOH to a Cabazitaxel derivative (CTX-C8-NH2 (SI-3)) (Fig. 1). 1,4-Dioxane– water titration of TMR-LTX probes yielded bell-shaped dose–
response curves, indicating two transitions corresponding to zwitterion formation followed by aggregation (ESI Fig. 4†). These processes are described by probeD50 and probeA50 parameters, respectively (ESI Table 3†). If the same titration is per-
formed keeping the SDS concentration constant at 0.3%, the aggregates cannot form and the data can be tted to a simple EC50 equation (ESI Fig. 4 and Table 3†).
For 50- and 60-regioisomers probeD50 were only marginally higher than the dyeD50 of corresponding uorophores. probeD50 of 40-regioisomers was signicantly increased compared to the corresponding dyeD50 (ESI Tables 2 and 3†). This indicates that
the conversion of the carboxyl group to amide induces higher
spirolactone content at equilibrium and thus implies higher hydrophobicity of 40-probes (Fig. 2a).
Fig. 2 Neighbouring group effect in the fluorescent probes. (a) DyeD50 values of positional isomers of TMR-COOH and probeD50 of TMR-LTX. (b) DFT optimized geometries of a model 40-regioisomer fluorescent probe in spirolactone and zwitterion forms with a truncated linker and targeting moiety. (c) DFT calculated potential energy differences between the spirolactone and zwitterion of model 40/50/60-regioisomer probes in
1,4-dioxane and water environment. (d) Chemical shifts of the amide proton of TMR-LTX regioisomeric probes. (e) Comparison of the retention
times of TMR-LTX regioisomeric probes in HPLC analysis with an SB-C18 column under isocratic elution conditions (75 : 25 MeOH : H2O, 25 mM HCOONH4, pH ¼ 3.6).
This journal is © The Royal Society of Chemistry 2020
Chem. Sci., 2020, 11, 7313–7323 | 7315
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Chemical Science
View Article Online
Edge Article
In order to better understand how NGE alters the spirolactone–zwitterion equilibrium, we modelled in silico 40/50/60carboxamidetetramethylrhodamine methylamides (Fig. 2b and ESI Fig. 5†) with a truncated ligand and linker. Geometrical structure optimizations pointed out unique properties of the 40isomer: the spirolactone form can form an intramolecular Hbond since the carboxamide proton and carbonyl are only 1.75 A apart; the carboxylate is 68 twisted with respect to the benzene ring because of steric hindrance in the zwitterion form (Fig. 2b and ESI Fig. 5†). Comparison of calculated total potential energies of the molecules in the modelled solvent environment representing two uttermost points in 1,4-dioxane (3r ¼ 2.21) and water (3r ¼ 78.35) indicates that the spirolactone form of 4-TMR-NHMe in 1,4-dioxane is stabilized and the zwitterion form in water is destabilized in comparison with other regioisomers (Fig. 2c). As the stabilization of the spirolactone form could be attributed to the formation of intramolecular hydrogen bonds, the stabilizing effect could be considered diminishing upon increase of water content. Therefore the observed shi of D50 value in 40-carboxamide isomers mostly arises from the destabilization of the zwitterion form in the water environment (ESI Table 4†). NMR spectroscopy conrmed the presence of the NGE: the amide proton in the 40-isomer probes is deshielded and downeld shied by 0.3–0.4 ppm compared to 50- and 60-regioisomers. This shi does not depend on the xanthene ring system or on the attached targeting ligand (Fig. 2d and ESI Table 5†).
The higher probeD50 values of 40-regioisomers suggest a higher percentage of the hydrophobic spirolactone form under equilibrium conditions which is retained longer in the reverse phase C18 HPLC column because of the stronger interaction with the resin.16 We compared retention times of the probes under isocratic elution conditions and observed a high correlation between probeD50 and the retention times (ESI Fig. 6†). The 40-regioisomer derivatives displayed signicantly longer elution times compared to 50- and 60-regioisomer based probes (Fig. 2e and ESI Table 6†).
Altogether, our experiments demonstrated that the NGE is present in the 40-isomer probes and encouraged us to test how it inuences the staining performance in living cells.
Characterization of probe–target interaction in vitro
Rhodamine probes are prone to aggregation, which signicantly contributes to their uorogenicity and might impair interaction with the target.4a We observed a high correlation between uorescence increase aer binding to tubulin and aer SDS addition, which indicates that affinity towards tubulin is sufficient to dissociate the probe from the aggregates (ESI Fig. 7†). In both cases, uorogenicity changes in the order TMR < 580CP z 610CP < SiR. Next, we carried out a tubulin polymerization assay, which showed that all probes were able to stabilize microtubules and are functional (ESI Fig. 8†).
Performance of 40-carboxyrhodamine tubulin probes in living cells
Cabazitaxel and Larotaxel are anticancer drugs displaying superior resistance to efflux pumps and brain blood
permeability.17 The cytotoxicity of tubulin probes is implied by the design, which exploits the attachment of uorophores to the 30-position of taxanes.4a We hypothesized that NGE enhanced spirolactonization should increase the cell membrane permeability, resulting in stronger staining of tubulin and higher cytotoxicity. Thus, we stained HeLa, U-2 OS and primary human dermal broblasts with the series of tubulin probes at 100 nM concentration (Fig. 3a and ESI Fig. 9†). All probes based on 40carboxamide dyes, with the exception of 4-SiR-CTX, stained the cells much more strongly compared to the 50- or 60-regioisomer. The improvement of staining ranged from 3- up to 20-fold for primary human broblasts (Fig. 3b). In good agreement with these results, all 40-regioisomer probes, except 4-SiR-CTX, demonstrated higher cytotoxicity in HeLa cells compared to the 50- and 60-regioisomer probes (Fig. 3c, ESI Fig. 10 and Table 7†).
Not only poor membrane permeability, but also active efflux might impede cell staining. To test the effect of multidrugresistance (MDR) efflux pumps, we took advantage of U-2 OS cells, which are known to express MDR proteins susceptible to a broad specicity inhibitor Verapamil.18 Staining with 4-TMRLTX, 4-580CP-LTX and 4-610CP-CTX was independent of Verapamil, suggesting that these probes are not good substrates of MDR efflux pumps. In contrast, the best performing siliconrhodamine probe 6-SiR-CTX was prone to the action of efflux pumps and required Verapamil for efficient staining (ESI Fig. 11†).
Probes targeting DNA and actin
Superior cell membrane permeability of NGE carrying tubulin probes encouraged us to test the versatility of this approach with the probes targeting DNA and actin. To this end, we synthesized 4-TMR-Hoechst, 4-580CP-Hoechst and 4-610CPJAS. All these probes stained their targets more intensively compared to 50- and 60-regioisomers in living cells (Fig. 4, ESI Fig. 12 and 13†).19 DNA probes showed similar affinity to previously reported 50- and 60-regioisomers and a single DNA binding mode similar to the 50-regioisomer (ESI Fig. 12b†).19b
Hoechst and its derivatives interfere with DNA synthesis leading to changes in S and G2/M phases.19b Jasplakinolide is known to cause multinucleation and apoptosis causing DNA fragmentation.20 Knowing these phenotypes, we monitored alterations in the cell cycle progression and identied the cytotoxicity threshold (ESI Fig. 12d and 13c†).
Confocal and Airyscan microscopy of living cells
Confocal images of tubulin stained with 40-regioisomers in living cells showed excellent contrast exceeding 50-fold single microtubule to cytoplasm signal ratio even if excess of the probe was not removed (Fig. 4a and ESI Fig. 14†). Airyscan images beneted from increased resolution ($150 nm) and reached up to 100-fold single microtubule to cytoplasm signal ratio (Fig. 4a and ESI Fig. 15†). This clearly demonstrates excellent permeability, high affinity and selectivity of the taxane probes. A similar improvement was observed for DNA and actin probes (Fig. 4, ESI Fig. 16 and movie 1†).
7316 | Chem. Sci., 2020, 11, 7313–7323
This journal is © The Royal Society of Chemistry 2020
Edge Article
View Article Online
Chemical Science
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Fig. 3 Imaging performance of fluorescent probes based on rhodamine isomers. (a) Wide-field fluorescence microscopy of living primary fibroblasts stained with 100 nM TMR-LTX isomers for 1 h at 37 C. Cells were washed once with HBSS and imaged in DMEM growth media. Inset shows zoomed-in images. Scale bars: 100 mm (large field of view), 10 mm (inset). Hoechst staining is shown in cyan and all tubulin probes are in magenta. (b) Quantification of fluorescence signal in the cytoplasm of living cells stained with tubulin probes. Data are presented as mean Æ s.e.m., N ¼ 3 independent experiments, each time n > 100 cells were quantified. (c) Cytotoxicity of tubulin fluorescent probes presented as half maximal effective concentration (EC50) after 24 h incubation at 37 C in growth media. Cytotoxicity was determined as the fraction of cells containing less than a single set of genetic material (sub G1 DNA content). Data are presented as mean Æ s.e.m., N ¼ 3 independent experiments, each time n > 100 cells were quantified. (d) Wide-field microscopy images of living primary fibroblasts. The cells were stained with 100 nM 4/5/6580CP-Hoechst (magenta) or 100 nM 4/5/6-610CP-JAS (yellow) for 1 h at 37 C, washed once with HBSS and imaged in DMEM media. Insets
show zoomed-in images. Scale bars: 100 mm (large field of view), 10 mm (insets). Overlay with phase contrast (grey) images are shown. (e) Quantification of DNA probe fluorescence signal in the nuclei. Data are presented as mean Æ s.d., N ¼ 3 independent experiments, each n > 100 cells. (f) Quantification of 4/5/6-610CP-JAS fluorescence signal in the cytoplasm of living cells. Data are presented as mean Æ s.d., N ¼ 3 independent experiments, each n > 100 cells.
The 4-TMR-LTX probe shows a very low cytotoxicity threshold (EC50 ¼ 6.3 Æ 0.5 nM, Fig. 3c), approaching the cytotoxicity of taxanes, which ranges from 1 to 10 nM21 (Table 1).
Taxanes are known to accumulate inside the cells up to 1000fold.22 Indeed, we were able to image broblasts stained with sub-nanomolar concentrations of 4-TMR-LTX (Fig. 4b). Despite
This journal is © The Royal Society of Chemistry 2020
Chem. Sci., 2020, 11, 7313–7323 | 7317
Chemical Science
View Article Online
Edge Article
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Fig. 4 Confocal and Airyscan imaging of living cells stained with rhodamine 40-isomer probes. (a) Single microtubule and cytoplasm fluorescence signal ratio in living human fibroblasts stained with 100 nM of the indicated probe for 1 h at 37 C and imaged without probe removal. Data are presented as mean Æ s.d., N $ 3 independent fields of view, each time n $ 20 microtubules. (b) Zeiss Airyscan images of human primary fibroblasts stained with 5-SiR-Hoechst and 4-TMR-LTX for 24 h at 37 C in growth media at indicated concentrations. Images acquired without probe removal. (c) Cell cycle of human primary fibroblasts stained with 1 nM 4-TMR-LTX (yellow) and 10 nM 5-610CP-Hoechst (magenta). Scale bar: 10 mm. Overlay with phase contrast (grey) images are shown. (d) Three-colour ZEISS Airyscan image of living HeLa cells at the metaphase stained with 3 nM 4-TMR-LTX (yellow), 20 nM 5-SiR-Hoechst (magenta) and 1000 nM 6-510R-JAS (cyan). Scale bar: 1 mm.
its relatively low uorogenicity, a washing step was not required to achieve high contrast. It is worth noting that at picomolar concentrations 4-TMR-LTX selectively stains the centrosome, likely due to a higher local tubulin concentration in this organelle (Fig. 4b, and ESI movie 2†).
The spectral separation between TMR and 610CP allows two colour sequential imaging (ESI Fig. 17†). We could resolve all
cell cycle stages in 3D while imaging HeLa cells stained with 4TMR-LTX and 5-610CP-Hoechst (Fig. 4c and ESI movie 3†). We exploited 6-510R-JAS,19a,23 4-TMR-LTX and 5-SiR-Hoechst (ref.
Table 1 Properties of the best performing probes
Probe
lambasx (nm)
lemmax (nm)
3 Â 103 (MÀ1 cmÀ1)
Tubulin probes
4-TMR-LTX
557
576
65 Æ 14
4-580CP-LTX
591
612
72 Æ 11
4-610CP-CTX
617
638
99 Æ 10
DNA probes
4-TMR-Hoechst
558
576
34 Æ 3
4-580CP-Hoechst 589
610
67 Æ 4
Actin probe
4-610CP-JAS
618
636
83 Æ 4
a Relative quantum yield. b Absolute quantum yield.
QY
0.56 Æ 0.02a 0.63 Æ 0.01a 0.73 Æ 0.02a
0.27 Æ 0.01b 0.51 Æ 0.01b
0.67 Æ 0.04a
ISsaTtED (MW cmÀ1)
4.91 Æ 0.69 2.29 Æ 0.14 2.75 Æ 0.28
— —
—
Brightness (MÀ1 cmÀ1)
36 100 Æ 8100 45 400 Æ 6700 72 000 Æ 7100
9100 Æ 900 34 400 Æ 1800
55 300 Æ 2500
Fltarget increase (fold)
2.3 Æ 0.4 13 Æ 3
8Æ1
87 Æ 7 34 Æ 1
40 Æ 3
Toxicity, EC50 (nM)
6.3 Æ 0.5 29 Æ 3 14 Æ 1
>4000 >1000
>500
7318 | Chem. Sci., 2020, 11, 7313–7323
This journal is © The Royal Society of Chemistry 2020
Edge Article
View Article Online
Chemical Science
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Fig. 5 STED nanoscopy imaging of living cells stained with rhodamine 40-isomer probes. (a) Confocal and STED images of microtubules in human fibroblasts taken with 0.25 AU pinhole. The cells were stained with 100 nM 4-610CP-CTX for 1 h at 37 C. Scale bar: 1 mm. (b) Fluorescence intensity profile at the rectangle in panel a. Insets show measured (mean Æ s.d., N ¼ 20) and predicted diameter from the cryo-electron
microscopy model of tubulin (orange) bound to paclitaxel (blue). (c) Apparent microtubule FWHM measured by different microscopy methods. Human fibroblasts stained with 100 nM probes for 1 h at 37 C. * – diameter measured between two peaks of the fitted intensity profile. Data
presented as mean Æ s.d., N $ 3 independent fields of view, each time n $ 10 microtubules. (d) Nine-fold symmetry of centriole resolved in the deconvolved STED DyMIN25 image of U-2 OS cell stained with 1 mM 4-610CP-CTX for 1 h at 37 C. White dashed lines mark a second centriole.
Scale bar: 1 mm. (e) Deconvolved STED image of human female fibroblast nucleus stained with 100 nM 4-580CP-Hoechst showing the inac-
tivated X chromosome (Xi). Insets – zoomed-in confocal and STED images of the Xi region. Scale bars: 500 nm (insets), 1 mm (main image). (f) Two-colour STED no-wash image of human fibroblasts stained with 100 nM 4-610CP-JAS and 10 nM 4-TMR-LTX for 1 h at 37 C. Scale bar: 10
mm.
19b) for three colour sequential Airyscan Z-stack imaging of mitotic HeLa cells meeting Nyquist criteria (50–50–150 nm in x– y–z axes) without signicant bleaching (Fig. 4d and ESI movie 4†).
Super-resolution STED microscopy of living cells
Finally, we applied our new probes in STED nanoscopy. The uorescence of TMR based probes could be inhibited by 660 and 775 nm lasers resulting in signicant improvement of the resolution and yielding apparent microtubule diameter below 80 nm in all the tested cases (ESI Fig. 18†). Despite signicant uorescence quenching by the STED laser (10 MW cmÀ2), we could reach up to 45-fold single microtubule to cytoplasm signal ratio while staining with 4-610CP-CTX (ESI Fig. 19† and 4a). We took advantage of the no-wash conditions and recorded STED images with up to a 100 frames time-course without signicant bleaching (ESI movie 5†). Good labelling efficiency and brightness of the probes allowed acquisition of STED images with 0.25 Airy units pinhole at increased resolution
(Fig. 5a–c). For the rst time we could resolve the microtubule diameter of 23 Æ 4 nm in living cells. This value closely matches the dimensions measured with cryo-electron microscopy (Fig. 5b).24 In addition, we obtained excellent quality images of tubulin in the centrosome demonstrating the 9-fold symmetry of this organelle (Fig. 5d). 3D STED allowed us to resolve, independent of orientation, the tube-like structure of centrioles stained with 4-610CP-CTX in living cells without probe removal (ESI Fig. 20 and movie 6†). Using the 4-580CP-Hoechst probe we visualized chromatin domain clusters of the inactivated X chromosome in primary human broblasts (Fig. 5e). Finally, we demonstrate two-colour STED imaging using 4-TMR-LTX and 4610CP-JAS probes (Fig. 5f and ESI Fig. 17d†).
Conclusions
Due to their excellent photophysical properties and biocompatibility, rhodamines are among the most popular dyes used for synthesis of uorescent probes for live-cell imaging.
This journal is © The Royal Society of Chemistry 2020
Chem. Sci., 2020, 11, 7313–7323 | 7319
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Chemical Science
View Article Online
Edge Article
Controlling the spirolactonization of rhodamine dyes facilitates cell permeability and increases uorogenicity of the probes. Herein we exploited the neighbouring group effect (NGE) occurring in the carboxyrhodamines' positional 40-conjugates for enhancing spirolactonization without any modications of the dye's core structure. This led to signicantly enhanced cell permeability, demonstrating a signicant contribution of the uorophore conguration to the properties of the probe.
Several studies have used dyeD50, the ability to form spirolactones, to guide rational uorophore design.5b,c,6,26 In our
case this would not have worked; with the exception of 4-SiRCOOH, dyeD50 values of 40-, 50- and 60-isomers are very close to each other, thus implying no major improvement at the free uorophore level. In contrast to free dyes, probes are prone to aggregation which can be estimated from uorescence increase
upon addition of a detergent (SDS) that dissolves aggregates (ESI Fig. 4 and 7†). Aer conjugation to the targeting ligand, one
of the neighbouring carboxylic groups is replaced with the
carboxamide group, which enhances spirolactonization and considerably increases probeD50 of 40-isomers. The spirolactone contributes to high contrast imaging by facilitating the probe's entry into the cell and decreasing the uorescence of the
unbound probe. It is worth noting that the latter mechanism is benecial but not essential for the high contrast imaging. If a probe can readily cross the cell membrane, it can be efficiently concentrated on the target structure due to the high affinity binding. For example, 1 nM 4-TMR-LTX is sufficient for efficient
tubulin staining (Fig. 4a and b). At such low concentrations, no washing step is required despite the low uorogenicity of 4-
TMR-LTX. Furthermore, at picomolar concentrations 4-TMR-
LTX selectively stains the centrosome. The unprecedented
permeability of the new tubulin probes allowed imaging at nanomolar–picomolar concentrations, which is in sharp
contrast to micromolar-high concentrations reported in other studies.5b,c,6,26,27
The high cell permeability of 4-610CP-CTX allowed exceptionally dense labelling of microtubules and provided suffi-
ciently strong signals for STED imaging with the pinhole set to
an extreme value of 0.25 AU. Thereby, to the best our knowledge, for the rst time, we resolved microtubule diameters of 23 Æ
4 nm in living cells, which matches the value obtained by cryoEM (Fig. 5a–c).24
NGE is benecial for probes based on more hydrophilic (low probeD50) 4-TMR-COOH, 4-580CP-COOH and 4-610CP-COOH dyes, but not on a more hydrophobic 4-SiR-COOH. In 4-SiRCTX, NGE pushes probeD50 outside the water polarity range, causing dominance of a non-absorbing spirolactone, which
results in decreased solubility and increased aggregation in water, which competes with the target binding (ESI Table 3†). This manifests in low background uorescence, but also in
weak polymerized tubulin stabilization in vitro, low cytotoxicity
and poor staining. The delicate balance between spirolactonization–aggregation and target binding is close to ideal for 6-SiRCTX, the best performing SiR-based probe, which has probeD50 close to the dielectric constant value of water (ESI Table 3†).
We demonstrated the generality of the NGE cell permeability
enhancement by synthesizing probes targeting DNA and actin,
which can also operate at low nanomolar concentrations indicating unrestricted permeability, very high affinity and specicity of the new probes. This decreases the chances for unspecic probe accumulation in the membranes. Although invisible, such off-targeting might disturb physiological processes.
In conclusion, we employed the neighbouring group effect in the 40-regioisomer of rhodamine uorophores to create, to the best of our knowledge, the most efficient probes for tubulin, actin and DNA labelling in living cells (Table 1 and ESI Table 7†). Their excellent spectroscopic properties, outstanding cell permeability and specicity make them powerful tools for many types of microscopy techniques, including super-resolution methods. This simple approach of positional isomerism is applicable to rhodamine- and uorescein-type uorophores, does not require modication of the xanthene ring system and can be used for generating new probes and for improving the performance of existing ones.
Experimental
Measurements of absorbance spectra in 1,4-dioxane–water mixtures
Measurements of the absorbance changes in 1,4-dioxane–water
mixtures were performed by pipetting 2 mL stock solutions of
dyes or probes in DMSO into a 96 glass bottom well plate (11
wells per sample) made from propylene (Corning 3364). To the
wells going from right to le 300 mL of 1,4-dioxane–water
mixtures containing 100%, 90%, 80%, 70%, 60%, 50%, 40%,
30%, 20%, 10% or 0% 1,4-dioxane was added (if needed,
mixtures with 0.3% SDS are used, with an exception in 100%
dioxane due to solubility issues). Aer incubation for 1 hour at
room temperature, absorption of solutions in each well was
recorded from 320 nm to 850 nm with wavelength step size of
1 nm on a multiwell plate reader Spark® 20 M (Tecan). The
background absorption of the glass bottom plate was measured
in wells containing only the 1,4-dioxane–water mixture with
a similar amount of DMSO and subtracted from the spectra of
the samples. 3max versus dielectric constants (D) of 1,4-dioxane–
water mixtures were plotted.15 D50 value was obtained by tting
to dose–response equation EC50 as implemented in GraphPad
6.0 soware:
,
Hill!
A ¼ A0 þ ðAmax À A0Þ 1 þ D50
(1)
d
where A0 – absorbance at 3max at 3r ¼ 0, Amax – the highest reached absorbance at 3max, d – dielectric constant of 1,4dioxane–water mixture at a given point, Hill – Hill slope coefcient determining the steepness of a dose–response curve, D50 – corresponds to the d value that provokes half of the absorbance amplitude (Amax À A0).
Data points from the 4-TMR-LTX titration were tted to the bell-shaped dose–response curve described by the following
equation:
A ¼ A0 þ
Amax À A0
À
Amax À A0
(2)
D50 Hill1
A50 Hill2
1þ d
1þ d
7320 | Chem. Sci., 2020, 11, 7313–7323
This journal is © The Royal Society of Chemistry 2020
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Edge Article
View Article Online
Chemical Science
where A0 – absorbance at 3max at 3r ¼ 0, Amax – the highest reached absorbance at 3max during the titration experiment. d – dielectric constant of 1,4-dioxane–water mixture at a given point, Hill1 – Hill slope coefficient determining the steepness of the ascending dose–response curve part, Hill2 – Hill slope coefficient determining the steepness of the declining dose– response curve part, D50 – corresponds to the d value that provokes half of the absorbance amplitude (Amax À A0) in the ascending dose–response curve part, A50 – corresponds to the d value that provokes half of the absorbance amplitude (Amax À A0) in the declining dose–response curve part.
Quantum chemical calculations
The initial geometries of the studied molecules were generated by using a molecular mechanics method (force eld MMF94, steepest descent algorithm) and a systematic conformational analysis as implemented in Avogadro 1.1.1 soware. The minimum energy conformer geometries found by molecular mechanics were further optimized with the Gaussian 09 program package28 by means of density functional theory (DFT) using the Becke, 3-parameter, Lee–Yang–Parr (B3LYP)29 exchange-correlation hybrid functional with 6-311G++(d,p) basis set,30 including the polarizable continuum model31 to simulate the environment in 1,4-dioxane or water. Furthermore, harmonic vibrational frequencies were calculated to verify the stability of the optimized geometries. All the calculated vibrational frequencies were real (positive) indicating the true minimum of the calculated total potential energy of the optimized system. The computations were performed at the High Performance Computing Center in G¨ottingen provided by GWDG.
Maintenance and preparation of cells
Human primary dermal broblasts were cultured in highglucose DMEM (Thermo Fisher, #31053044) and 10% FBS (Thermo Fisher, #10082147) with the presence of 1% penicillin– streptomycin (Sigma, #P0781) in a humidied 5% CO2 incubator at 37 C. HeLa cells were cultured in high-glucose DMEM (Thermo Fisher, #31966047) and 10% FBS (Bio&SELL #S0615), 1 mM sodium pyruvate (Sigma, #S8636) with the presence of 1% penicillin–streptomycin (Sigma, #P0781) in a humidied 5% CO2 incubator at 37 C. The cells were split every 3–4 days or at conuence. Cells were seeded in glass bottom 12-well plates (MatTek, #P12G-1.0-14-F) before imaging experiments. The U2OS cells were cultured in McCoys 5A medium (Thermo sher #16600082) and 10% FBS (Bio&SELL, #S0615) with the presence of 1 mM sodium pyruvate (Sigma, #S8636) and 1% of penicillin– streptomycin (Sigma #P0781) in a humidied 5% CO2 incubator at 37 C.
Cells were stained with the uorescent probes in DMEM (Thermo Fisher, #31053044) supplemented with 10% FBS ((Bio&SELL, #S0615)) at 37 C and 5% CO2. If needed, the cells were washed 2 times with HBSS (Lonza, #BE10-527F) and imaged in DMEM with 10% FBS. No-wash experiments were performed in DMEM with 10% FBS aer probe addition and incubation for the indicated period of time. For Airyscan
experiments, cells were seeded in a 10-well plate (Greiner bioone, #543079) 1 day before staining. The probes were applied to the cells in DMEM medium, and aer 1 h or overnight incubation cells were imaged without washing.
STED microscope with a 775 nm laser
Comparative confocal and STED images were acquired on an Abberior STED 775 QUAD scanning microscope (Abberior Instruments GmbH) equipped with 561 nm and 640 nm 40 MHz pulsed excitation lasers, a pulsed 775 nm 40 MHz STED laser, and an UPlanSApo 100Â/1.40 Oil objective. The following detection windows were used: for the TMR/580CP channel 615/20 nm, and for the 610CP/SiR channel 685/70 nm. In this setup, voxel size was 15–30 nm in the xy plane and 150 nm in the z-axis for 2D STED images. For the STED imaging with a 0.25 AU pinhole, the pixel size was set to 7 Â 7 nm and images were acquired with 2 line accumulation. Laser powers were optimized for each sample. 3D STED images were acquired using a pinhole set to 0.8 AU, voxel size set to 40 Â 40 Â 40 nm, 3D STED doughnut set to 90%, with single line accumulation and xzy scanning mode.
Estimation of the STED effect was performed by varying the STED laser power from 0 to 100% while measuring cells stained with tubulin probes. Obtained data were tted using GraphPad Prism 6 to the following equation:
Y ¼ pffiffiffidffiffifficffioffiffinffifffiffiffiffiffiffiffiffiffi
(3)
1 þ I =Isat
where dconf – confocal resolution, I – STED laser intensity power, Isat – saturating STED laser intensity power.
STED microscope with a 660 nm laser
STED images with a 660 nm STED laser were acquired on a Leica TCS SP8 STED 3X scanning microscope (Leica Microsystems) equipped with an HC PL APO CS2 100Â/1.40 oil objective and the incubator set to 37 C and 5% CO2. The 4-TMR-LTX probe was excited with the 561 nm, 80 MHz pulsed line of a white light laser, de-excited with a continuous wave 660 nm STED laser and detected in the 600/50 nm window. In this setup, the pixel size was set to 23 nm in the xy plane. The STED effect was estimated as described above.
Airyscan microscope
Live-cell imaging was performed on a Zeiss LSM880 system (Carl Zeiss) equipped with an oil immersion Plan-Apochromat 63Â/1.40 Oil Corr M27 objective and the incubator set to 37 C and 5% CO2. The following excitation lasers in combination with confocal or Airyscan acquisition regimes were used: argon laser 488 nm for 510R probes, 561 nm diode laser for TMR and 580CP probes, and 633 nm diode laser for 610CP and SiR probes. For TMR and 610CP, a combination of BP 570–620 + LP 645 lters with dichroic beam splitter SBS SP 615 in the TMR channel was used. The TMR, 510R and SiR combination was excited and scanned in sequence using BP 495–550 + LP 570 lters. Laser powers were optimized for each sample. Detailed excitation and detection schemes are shown in ESI Fig. 16.†
This journal is © The Royal Society of Chemistry 2020
Chem. Sci., 2020, 11, 7313–7323 | 7321
Enhancing the biocompatibility of rhodamine fluorescent probes by a neighbouring group effect
Our study demonstrates that the neighbouring group effect which results from positioning an amide group next to a carboxyl group in the benzene ring of rhodamines dramatically increases cell permeability of the rhodaminebased probes through stabilizing a fluorophore in a hydrophobic spirolactone state. Based on this principle, we create probes targeting tubulin, actin and DNA. Their superb staining efficiency allows long-term confocal and 3D Airyscan microscopy, and stimulated emission depletion (STED) nanoscopy with sub-30 nm resolution in living cells. Artwork prepared by Hartmut Sebesse and Gražvydas Lukinavičius.
As featured in:
See Gražvydas Lukinavičius et al., Chem. Sci., 2020, 11, 7313.
rsc.li/chemical-science
Registered charity number: 207890
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Chemical Science
EDGE ARTICLE
View Article Online
View Journal | View Issue
Cite this: Chem. Sci., 2020, 11, 7313 All publication charges for this article
have been paid for by the Royal Society of Chemistry
Received 15th April 2020 Accepted 20th June 2020 DOI: 10.1039/d0sc02154g rsc.li/chemical-science
Enhancing the biocompatibility of rhodamine fluorescent probes by a neighbouring group effect†
Jonas Bucevicˇius, a Georgij Kostiuk, and Graˇzvydas Lukinavicˇius *a
a Ru¯ta Gerasimaite˙,
a Tanja Gilatb
Fluorescence microscopy is an essential tool for understanding dynamic processes in living cells and organisms. However, many fluorescent probes for labelling cellular structures suffer from unspecific interactions and low cell permeability. Herein, we demonstrate that the neighbouring group effect which results from positioning an amide group next to a carboxyl group in the benzene ring of rhodamines dramatically increases cell permeability of the rhodamine-based probes through stabilizing a fluorophore in a hydrophobic spirolactone state. Based on this principle, we create probes targeting tubulin, actin and DNA. Their superb staining intensity, tuned toxicity and specificity allows long-term 3D confocal and STED nanoscopy with sub-30 nm resolution. Due to their unrestricted cell permeability and efficient accumulation on the target, the new probes produce high contrast images at low nanomolar concentrations. Superior performance is exemplified by resolving the real microtubule diameter of 23 nm and selective staining of the centrosome inside living cells for the first time.
Introduction
Modern super-resolution uorescence microscopy and nanoscopy techniques permit observation of biological processes in living organisms down to the molecular level.1 However, their success depends on the availability of biocompatible uorescent probes for specic labelling of cellular structures. Usually, these probes are constructed by coupling a uorescent dye via a linker to a small molecule ligand targeting a protein of interest. The linker and the uorophore make signicant contributions to the nal properties of such probes, oen resulting in compromised cell permeability and off-targeting.2 Recent studies dedicate a lot of attention to the rhodamine class uorophores that can cycle between a non-uorescent spirolactone and a uorescent zwitterion form. This property is exploited to generate uorogenic probes, which are sensitive to a number of environmental factors: pH, ion concentration, enzymatic activity, local microenvironment polarity or light.3 Since the spirolactone is a more hydrophobic molecule compared to the zwitterion, it makes rhodamine-based uorescent probes more cell-permeable.4 Regular rhodamines and carbon-substituted analogues (carbopyronines) have the
aChromatin Labeling and Imaging Group, Department of NanoBiophotonics, Max Planck Institute for Biophysical Chemistry, Am Fassberg 11, 37077 G¨ottingen, Germany. E-mail: [email protected] bDepartment of NanoBiophotonics, Max Planck Institute for Biophysical Chemistry, Am Fassberg 11, 37077 G¨ottingen, Germany † Electronic supplementary information (ESI) available. See DOI: 10.1039/d0sc02154g
equilibrium shied towards the zwitterion form, which results in relatively poor membrane permeability.4 To overcome this issue, several studies attempted to induce spirolactone form preference by introducing electron-withdrawing groups into the xanthene core5 or in the benzoic acid substituent.6 However, these approaches result in a bulkier core structure and alter the physicochemical properties of the dyes. In addition, all of them must rely on the uorogenicity in order to minimize background staining arising from the high working concentrations of the probes. We propose a way for modulating the spirolactone–zwitterion equilibrium without any change in the uorescent dye core structure by exploiting the neighbouring group effect (NGE), i.e. the phenomenon that two neighbouring carboxylic groups can inuence each other via steric, electrostatic or H-bond interactions.7 The NGE should not be confused with the term “neighbouring group participation” (NGP) – the direct interaction with a reactive center during a chemical reaction.8
Here, we introduce a new positional isomer with NGE to increase the cell permeability and uorogenicity of regular rhodamines without changing their spectroscopic properties. We demonstrate that excellent quality of wash-free multi-colour live-cell images can be obtained at low working concentrations with wide-eld, confocal and super-resolution microscopes even with moderately uorogenic probes. Selective centrosome staining is demonstrated using exceptionally well performing tubulin probes. Superior performance is exemplied by resolving the real microtubule diameter of 23 nm in STED nanoscopy images.
This journal is © The Royal Society of Chemistry 2020
Chem. Sci., 2020, 11, 7313–7323 | 7313
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Chemical Science
View Article Online
Edge Article
Results and discussion
Chemical synthesis of 40-carboxyl rhodamines
For a long time, the existence of a novel isomeric class of 40carboxyrhodamines was debated due to the synthesis challenge arising from steric hindrance and the ortho-effect altered reactivity of adjacent carbonyl groups.9 Isomeric mixtures of 50/ 60-carboxyrhodamines and uoresceins are usually obtained by the classical condensation process of phthalic anhydrides and 3-dimethylaminophenol or resorcinol respectively.10
The condensation between resorcinol and 1,2,3-benzenetricarboxylic acid or its homologs yields solely 70-carboxyuorescein and not 40-carboxyuorescein.11 Another widely applied and regioselective synthesis strategy, which can be applied in the synthesis of 40-carboxyrhodamines, involves addition of aryl-lithium species to the 3,6-substitutedxanthone.12 We used formation of di-tert-butyl ester 1 for the protection of 3-bromophthalic acid carboxyl groups for the halogen–lithium exchange reaction (Fig. 1). However, the tertbutyl protecting group has low tolerance to strong nucleophiles and the lithium–halogen exchange must be carried out at
temperatures between À90 and À120 C. Such temperatures are
not compatible with the reactivity of xanthones and their C/Si
analogues possessing strong electron donating dialkyl amino
groups at positions 3 and 6. When the halogen exchange was attempted at À78 C, mainly homo-addition products were
formed. To solve this issue, we performed the reaction at $À116 C with the more reactive silyl protected 3,6-dihydrox-
yxanthone 2 and its C (3) and Si (4) analogues and performed the late stage uorescein to rhodamine conversion strategy.13 To achieve this, TBDMS protecting groups on compounds 5–7 were removed with TBAF in THF and the obtained uorescein class compounds 8–10 were reacted with triic anhydride in the presence of pyridine in DCM to obtain the triates 11–13. The triate groups were converted to the desired (di)methylamino groups (compounds 14–16) by a Pd-catalyzed C–N crosscoupling reaction and the desired 40-carboxyrhodamines were
obtained by quantitative deprotection with a 1 : 1 mixture of DCM/TFA (Fig. 1). The 40-carboxyrhodamines behave differently
than other regioisomers upon activation with common peptide coupling reagents – the formed carbonyl activated species (NHS
ester, HOBt ester or similar) immediately undergoes
Fig. 1 Synthesis scheme of 4-TMR-COOH, 4-580CP-COOH, 4-610CP-COOH and 4-SiR-COOH dyes and conjugates 22–42. Their photophysical properties are depicted in the colour spectrum chart and the neighbouring group effect is highlighted.
7314 | Chem. Sci., 2020, 11, 7313–7323
This journal is © The Royal Society of Chemistry 2020
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Edge Article
View Article Online
Chemical Science
intramolecular cyclisation forming the anhydride (Fig. 1 and ESI Fig. 1†). In the presence of amines as nucleophiles the major products form via nucleophilic attack on a less hindered carbonyl group (position 40), although traces of spiroamide products can be observed by LC/MS analysis.
In vitro characterization of the neighbouring group effect in 40-carboxyrhodamines
40-, 50- and 60-Carboxyrhodamine regioisomers have almost identical absorption, uorescence emission spectra and quantum yield (QY) in phosphate buffered saline (PBS) (ESI Table 1†). In agreement with previously published observations, we obtained an excellent correlation between QY and uorescence lifetime (ESI Fig. 2a†).14 Even better correlation was obtained between uorophore brightness (product of QY and extinction coefficient) and uorescence lifetime with the exception of the 4-SiR-COOH dye, which showed a signicantly lower extinction coefficient (ESI Fig. 2b†).
To characterize the ability of the uorophores to switch between spirolactone and zwitterion states, we measured the absorbance in water–1,4-dioxane mixtures with known dielectric constants.15 We tted the absorbance change to the dose– response equation and obtained the dyeD50 value, which indicates the dielectric constant at which absorbance is halved (ESI Table 2 and Fig. 3†). dyeD50 of 4-SiR-COOH was signicantly
higher than the dyeD50 of 50- and 60-isomers. For all other uorophores, the dyeD50 of all three isomers were very close.
We have previously demonstrated that the performance of
the tubulin probes depends on the chosen taxane ligand, and that the optimal ligand is different for each uorophore.4a
Based on this knowledge, we synthesized a new series of tubulin probes by coupling the 40-, 50- and 60-regioisomers of TMRCOOH and 580CP-COOH uorophores to a Larotaxel derivative
(LTX-C8-NH2 (SI-4)) and 610CP-COOH and SiR-COOH to a Cabazitaxel derivative (CTX-C8-NH2 (SI-3)) (Fig. 1). 1,4-Dioxane– water titration of TMR-LTX probes yielded bell-shaped dose–
response curves, indicating two transitions corresponding to zwitterion formation followed by aggregation (ESI Fig. 4†). These processes are described by probeD50 and probeA50 parameters, respectively (ESI Table 3†). If the same titration is per-
formed keeping the SDS concentration constant at 0.3%, the aggregates cannot form and the data can be tted to a simple EC50 equation (ESI Fig. 4 and Table 3†).
For 50- and 60-regioisomers probeD50 were only marginally higher than the dyeD50 of corresponding uorophores. probeD50 of 40-regioisomers was signicantly increased compared to the corresponding dyeD50 (ESI Tables 2 and 3†). This indicates that
the conversion of the carboxyl group to amide induces higher
spirolactone content at equilibrium and thus implies higher hydrophobicity of 40-probes (Fig. 2a).
Fig. 2 Neighbouring group effect in the fluorescent probes. (a) DyeD50 values of positional isomers of TMR-COOH and probeD50 of TMR-LTX. (b) DFT optimized geometries of a model 40-regioisomer fluorescent probe in spirolactone and zwitterion forms with a truncated linker and targeting moiety. (c) DFT calculated potential energy differences between the spirolactone and zwitterion of model 40/50/60-regioisomer probes in
1,4-dioxane and water environment. (d) Chemical shifts of the amide proton of TMR-LTX regioisomeric probes. (e) Comparison of the retention
times of TMR-LTX regioisomeric probes in HPLC analysis with an SB-C18 column under isocratic elution conditions (75 : 25 MeOH : H2O, 25 mM HCOONH4, pH ¼ 3.6).
This journal is © The Royal Society of Chemistry 2020
Chem. Sci., 2020, 11, 7313–7323 | 7315
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Chemical Science
View Article Online
Edge Article
In order to better understand how NGE alters the spirolactone–zwitterion equilibrium, we modelled in silico 40/50/60carboxamidetetramethylrhodamine methylamides (Fig. 2b and ESI Fig. 5†) with a truncated ligand and linker. Geometrical structure optimizations pointed out unique properties of the 40isomer: the spirolactone form can form an intramolecular Hbond since the carboxamide proton and carbonyl are only 1.75 A apart; the carboxylate is 68 twisted with respect to the benzene ring because of steric hindrance in the zwitterion form (Fig. 2b and ESI Fig. 5†). Comparison of calculated total potential energies of the molecules in the modelled solvent environment representing two uttermost points in 1,4-dioxane (3r ¼ 2.21) and water (3r ¼ 78.35) indicates that the spirolactone form of 4-TMR-NHMe in 1,4-dioxane is stabilized and the zwitterion form in water is destabilized in comparison with other regioisomers (Fig. 2c). As the stabilization of the spirolactone form could be attributed to the formation of intramolecular hydrogen bonds, the stabilizing effect could be considered diminishing upon increase of water content. Therefore the observed shi of D50 value in 40-carboxamide isomers mostly arises from the destabilization of the zwitterion form in the water environment (ESI Table 4†). NMR spectroscopy conrmed the presence of the NGE: the amide proton in the 40-isomer probes is deshielded and downeld shied by 0.3–0.4 ppm compared to 50- and 60-regioisomers. This shi does not depend on the xanthene ring system or on the attached targeting ligand (Fig. 2d and ESI Table 5†).
The higher probeD50 values of 40-regioisomers suggest a higher percentage of the hydrophobic spirolactone form under equilibrium conditions which is retained longer in the reverse phase C18 HPLC column because of the stronger interaction with the resin.16 We compared retention times of the probes under isocratic elution conditions and observed a high correlation between probeD50 and the retention times (ESI Fig. 6†). The 40-regioisomer derivatives displayed signicantly longer elution times compared to 50- and 60-regioisomer based probes (Fig. 2e and ESI Table 6†).
Altogether, our experiments demonstrated that the NGE is present in the 40-isomer probes and encouraged us to test how it inuences the staining performance in living cells.
Characterization of probe–target interaction in vitro
Rhodamine probes are prone to aggregation, which signicantly contributes to their uorogenicity and might impair interaction with the target.4a We observed a high correlation between uorescence increase aer binding to tubulin and aer SDS addition, which indicates that affinity towards tubulin is sufficient to dissociate the probe from the aggregates (ESI Fig. 7†). In both cases, uorogenicity changes in the order TMR < 580CP z 610CP < SiR. Next, we carried out a tubulin polymerization assay, which showed that all probes were able to stabilize microtubules and are functional (ESI Fig. 8†).
Performance of 40-carboxyrhodamine tubulin probes in living cells
Cabazitaxel and Larotaxel are anticancer drugs displaying superior resistance to efflux pumps and brain blood
permeability.17 The cytotoxicity of tubulin probes is implied by the design, which exploits the attachment of uorophores to the 30-position of taxanes.4a We hypothesized that NGE enhanced spirolactonization should increase the cell membrane permeability, resulting in stronger staining of tubulin and higher cytotoxicity. Thus, we stained HeLa, U-2 OS and primary human dermal broblasts with the series of tubulin probes at 100 nM concentration (Fig. 3a and ESI Fig. 9†). All probes based on 40carboxamide dyes, with the exception of 4-SiR-CTX, stained the cells much more strongly compared to the 50- or 60-regioisomer. The improvement of staining ranged from 3- up to 20-fold for primary human broblasts (Fig. 3b). In good agreement with these results, all 40-regioisomer probes, except 4-SiR-CTX, demonstrated higher cytotoxicity in HeLa cells compared to the 50- and 60-regioisomer probes (Fig. 3c, ESI Fig. 10 and Table 7†).
Not only poor membrane permeability, but also active efflux might impede cell staining. To test the effect of multidrugresistance (MDR) efflux pumps, we took advantage of U-2 OS cells, which are known to express MDR proteins susceptible to a broad specicity inhibitor Verapamil.18 Staining with 4-TMRLTX, 4-580CP-LTX and 4-610CP-CTX was independent of Verapamil, suggesting that these probes are not good substrates of MDR efflux pumps. In contrast, the best performing siliconrhodamine probe 6-SiR-CTX was prone to the action of efflux pumps and required Verapamil for efficient staining (ESI Fig. 11†).
Probes targeting DNA and actin
Superior cell membrane permeability of NGE carrying tubulin probes encouraged us to test the versatility of this approach with the probes targeting DNA and actin. To this end, we synthesized 4-TMR-Hoechst, 4-580CP-Hoechst and 4-610CPJAS. All these probes stained their targets more intensively compared to 50- and 60-regioisomers in living cells (Fig. 4, ESI Fig. 12 and 13†).19 DNA probes showed similar affinity to previously reported 50- and 60-regioisomers and a single DNA binding mode similar to the 50-regioisomer (ESI Fig. 12b†).19b
Hoechst and its derivatives interfere with DNA synthesis leading to changes in S and G2/M phases.19b Jasplakinolide is known to cause multinucleation and apoptosis causing DNA fragmentation.20 Knowing these phenotypes, we monitored alterations in the cell cycle progression and identied the cytotoxicity threshold (ESI Fig. 12d and 13c†).
Confocal and Airyscan microscopy of living cells
Confocal images of tubulin stained with 40-regioisomers in living cells showed excellent contrast exceeding 50-fold single microtubule to cytoplasm signal ratio even if excess of the probe was not removed (Fig. 4a and ESI Fig. 14†). Airyscan images beneted from increased resolution ($150 nm) and reached up to 100-fold single microtubule to cytoplasm signal ratio (Fig. 4a and ESI Fig. 15†). This clearly demonstrates excellent permeability, high affinity and selectivity of the taxane probes. A similar improvement was observed for DNA and actin probes (Fig. 4, ESI Fig. 16 and movie 1†).
7316 | Chem. Sci., 2020, 11, 7313–7323
This journal is © The Royal Society of Chemistry 2020
Edge Article
View Article Online
Chemical Science
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Fig. 3 Imaging performance of fluorescent probes based on rhodamine isomers. (a) Wide-field fluorescence microscopy of living primary fibroblasts stained with 100 nM TMR-LTX isomers for 1 h at 37 C. Cells were washed once with HBSS and imaged in DMEM growth media. Inset shows zoomed-in images. Scale bars: 100 mm (large field of view), 10 mm (inset). Hoechst staining is shown in cyan and all tubulin probes are in magenta. (b) Quantification of fluorescence signal in the cytoplasm of living cells stained with tubulin probes. Data are presented as mean Æ s.e.m., N ¼ 3 independent experiments, each time n > 100 cells were quantified. (c) Cytotoxicity of tubulin fluorescent probes presented as half maximal effective concentration (EC50) after 24 h incubation at 37 C in growth media. Cytotoxicity was determined as the fraction of cells containing less than a single set of genetic material (sub G1 DNA content). Data are presented as mean Æ s.e.m., N ¼ 3 independent experiments, each time n > 100 cells were quantified. (d) Wide-field microscopy images of living primary fibroblasts. The cells were stained with 100 nM 4/5/6580CP-Hoechst (magenta) or 100 nM 4/5/6-610CP-JAS (yellow) for 1 h at 37 C, washed once with HBSS and imaged in DMEM media. Insets
show zoomed-in images. Scale bars: 100 mm (large field of view), 10 mm (insets). Overlay with phase contrast (grey) images are shown. (e) Quantification of DNA probe fluorescence signal in the nuclei. Data are presented as mean Æ s.d., N ¼ 3 independent experiments, each n > 100 cells. (f) Quantification of 4/5/6-610CP-JAS fluorescence signal in the cytoplasm of living cells. Data are presented as mean Æ s.d., N ¼ 3 independent experiments, each n > 100 cells.
The 4-TMR-LTX probe shows a very low cytotoxicity threshold (EC50 ¼ 6.3 Æ 0.5 nM, Fig. 3c), approaching the cytotoxicity of taxanes, which ranges from 1 to 10 nM21 (Table 1).
Taxanes are known to accumulate inside the cells up to 1000fold.22 Indeed, we were able to image broblasts stained with sub-nanomolar concentrations of 4-TMR-LTX (Fig. 4b). Despite
This journal is © The Royal Society of Chemistry 2020
Chem. Sci., 2020, 11, 7313–7323 | 7317
Chemical Science
View Article Online
Edge Article
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Fig. 4 Confocal and Airyscan imaging of living cells stained with rhodamine 40-isomer probes. (a) Single microtubule and cytoplasm fluorescence signal ratio in living human fibroblasts stained with 100 nM of the indicated probe for 1 h at 37 C and imaged without probe removal. Data are presented as mean Æ s.d., N $ 3 independent fields of view, each time n $ 20 microtubules. (b) Zeiss Airyscan images of human primary fibroblasts stained with 5-SiR-Hoechst and 4-TMR-LTX for 24 h at 37 C in growth media at indicated concentrations. Images acquired without probe removal. (c) Cell cycle of human primary fibroblasts stained with 1 nM 4-TMR-LTX (yellow) and 10 nM 5-610CP-Hoechst (magenta). Scale bar: 10 mm. Overlay with phase contrast (grey) images are shown. (d) Three-colour ZEISS Airyscan image of living HeLa cells at the metaphase stained with 3 nM 4-TMR-LTX (yellow), 20 nM 5-SiR-Hoechst (magenta) and 1000 nM 6-510R-JAS (cyan). Scale bar: 1 mm.
its relatively low uorogenicity, a washing step was not required to achieve high contrast. It is worth noting that at picomolar concentrations 4-TMR-LTX selectively stains the centrosome, likely due to a higher local tubulin concentration in this organelle (Fig. 4b, and ESI movie 2†).
The spectral separation between TMR and 610CP allows two colour sequential imaging (ESI Fig. 17†). We could resolve all
cell cycle stages in 3D while imaging HeLa cells stained with 4TMR-LTX and 5-610CP-Hoechst (Fig. 4c and ESI movie 3†). We exploited 6-510R-JAS,19a,23 4-TMR-LTX and 5-SiR-Hoechst (ref.
Table 1 Properties of the best performing probes
Probe
lambasx (nm)
lemmax (nm)
3 Â 103 (MÀ1 cmÀ1)
Tubulin probes
4-TMR-LTX
557
576
65 Æ 14
4-580CP-LTX
591
612
72 Æ 11
4-610CP-CTX
617
638
99 Æ 10
DNA probes
4-TMR-Hoechst
558
576
34 Æ 3
4-580CP-Hoechst 589
610
67 Æ 4
Actin probe
4-610CP-JAS
618
636
83 Æ 4
a Relative quantum yield. b Absolute quantum yield.
QY
0.56 Æ 0.02a 0.63 Æ 0.01a 0.73 Æ 0.02a
0.27 Æ 0.01b 0.51 Æ 0.01b
0.67 Æ 0.04a
ISsaTtED (MW cmÀ1)
4.91 Æ 0.69 2.29 Æ 0.14 2.75 Æ 0.28
— —
—
Brightness (MÀ1 cmÀ1)
36 100 Æ 8100 45 400 Æ 6700 72 000 Æ 7100
9100 Æ 900 34 400 Æ 1800
55 300 Æ 2500
Fltarget increase (fold)
2.3 Æ 0.4 13 Æ 3
8Æ1
87 Æ 7 34 Æ 1
40 Æ 3
Toxicity, EC50 (nM)
6.3 Æ 0.5 29 Æ 3 14 Æ 1
>4000 >1000
>500
7318 | Chem. Sci., 2020, 11, 7313–7323
This journal is © The Royal Society of Chemistry 2020
Edge Article
View Article Online
Chemical Science
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Fig. 5 STED nanoscopy imaging of living cells stained with rhodamine 40-isomer probes. (a) Confocal and STED images of microtubules in human fibroblasts taken with 0.25 AU pinhole. The cells were stained with 100 nM 4-610CP-CTX for 1 h at 37 C. Scale bar: 1 mm. (b) Fluorescence intensity profile at the rectangle in panel a. Insets show measured (mean Æ s.d., N ¼ 20) and predicted diameter from the cryo-electron
microscopy model of tubulin (orange) bound to paclitaxel (blue). (c) Apparent microtubule FWHM measured by different microscopy methods. Human fibroblasts stained with 100 nM probes for 1 h at 37 C. * – diameter measured between two peaks of the fitted intensity profile. Data
presented as mean Æ s.d., N $ 3 independent fields of view, each time n $ 10 microtubules. (d) Nine-fold symmetry of centriole resolved in the deconvolved STED DyMIN25 image of U-2 OS cell stained with 1 mM 4-610CP-CTX for 1 h at 37 C. White dashed lines mark a second centriole.
Scale bar: 1 mm. (e) Deconvolved STED image of human female fibroblast nucleus stained with 100 nM 4-580CP-Hoechst showing the inac-
tivated X chromosome (Xi). Insets – zoomed-in confocal and STED images of the Xi region. Scale bars: 500 nm (insets), 1 mm (main image). (f) Two-colour STED no-wash image of human fibroblasts stained with 100 nM 4-610CP-JAS and 10 nM 4-TMR-LTX for 1 h at 37 C. Scale bar: 10
mm.
19b) for three colour sequential Airyscan Z-stack imaging of mitotic HeLa cells meeting Nyquist criteria (50–50–150 nm in x– y–z axes) without signicant bleaching (Fig. 4d and ESI movie 4†).
Super-resolution STED microscopy of living cells
Finally, we applied our new probes in STED nanoscopy. The uorescence of TMR based probes could be inhibited by 660 and 775 nm lasers resulting in signicant improvement of the resolution and yielding apparent microtubule diameter below 80 nm in all the tested cases (ESI Fig. 18†). Despite signicant uorescence quenching by the STED laser (10 MW cmÀ2), we could reach up to 45-fold single microtubule to cytoplasm signal ratio while staining with 4-610CP-CTX (ESI Fig. 19† and 4a). We took advantage of the no-wash conditions and recorded STED images with up to a 100 frames time-course without signicant bleaching (ESI movie 5†). Good labelling efficiency and brightness of the probes allowed acquisition of STED images with 0.25 Airy units pinhole at increased resolution
(Fig. 5a–c). For the rst time we could resolve the microtubule diameter of 23 Æ 4 nm in living cells. This value closely matches the dimensions measured with cryo-electron microscopy (Fig. 5b).24 In addition, we obtained excellent quality images of tubulin in the centrosome demonstrating the 9-fold symmetry of this organelle (Fig. 5d). 3D STED allowed us to resolve, independent of orientation, the tube-like structure of centrioles stained with 4-610CP-CTX in living cells without probe removal (ESI Fig. 20 and movie 6†). Using the 4-580CP-Hoechst probe we visualized chromatin domain clusters of the inactivated X chromosome in primary human broblasts (Fig. 5e). Finally, we demonstrate two-colour STED imaging using 4-TMR-LTX and 4610CP-JAS probes (Fig. 5f and ESI Fig. 17d†).
Conclusions
Due to their excellent photophysical properties and biocompatibility, rhodamines are among the most popular dyes used for synthesis of uorescent probes for live-cell imaging.
This journal is © The Royal Society of Chemistry 2020
Chem. Sci., 2020, 11, 7313–7323 | 7319
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Chemical Science
View Article Online
Edge Article
Controlling the spirolactonization of rhodamine dyes facilitates cell permeability and increases uorogenicity of the probes. Herein we exploited the neighbouring group effect (NGE) occurring in the carboxyrhodamines' positional 40-conjugates for enhancing spirolactonization without any modications of the dye's core structure. This led to signicantly enhanced cell permeability, demonstrating a signicant contribution of the uorophore conguration to the properties of the probe.
Several studies have used dyeD50, the ability to form spirolactones, to guide rational uorophore design.5b,c,6,26 In our
case this would not have worked; with the exception of 4-SiRCOOH, dyeD50 values of 40-, 50- and 60-isomers are very close to each other, thus implying no major improvement at the free uorophore level. In contrast to free dyes, probes are prone to aggregation which can be estimated from uorescence increase
upon addition of a detergent (SDS) that dissolves aggregates (ESI Fig. 4 and 7†). Aer conjugation to the targeting ligand, one
of the neighbouring carboxylic groups is replaced with the
carboxamide group, which enhances spirolactonization and considerably increases probeD50 of 40-isomers. The spirolactone contributes to high contrast imaging by facilitating the probe's entry into the cell and decreasing the uorescence of the
unbound probe. It is worth noting that the latter mechanism is benecial but not essential for the high contrast imaging. If a probe can readily cross the cell membrane, it can be efficiently concentrated on the target structure due to the high affinity binding. For example, 1 nM 4-TMR-LTX is sufficient for efficient
tubulin staining (Fig. 4a and b). At such low concentrations, no washing step is required despite the low uorogenicity of 4-
TMR-LTX. Furthermore, at picomolar concentrations 4-TMR-
LTX selectively stains the centrosome. The unprecedented
permeability of the new tubulin probes allowed imaging at nanomolar–picomolar concentrations, which is in sharp
contrast to micromolar-high concentrations reported in other studies.5b,c,6,26,27
The high cell permeability of 4-610CP-CTX allowed exceptionally dense labelling of microtubules and provided suffi-
ciently strong signals for STED imaging with the pinhole set to
an extreme value of 0.25 AU. Thereby, to the best our knowledge, for the rst time, we resolved microtubule diameters of 23 Æ
4 nm in living cells, which matches the value obtained by cryoEM (Fig. 5a–c).24
NGE is benecial for probes based on more hydrophilic (low probeD50) 4-TMR-COOH, 4-580CP-COOH and 4-610CP-COOH dyes, but not on a more hydrophobic 4-SiR-COOH. In 4-SiRCTX, NGE pushes probeD50 outside the water polarity range, causing dominance of a non-absorbing spirolactone, which
results in decreased solubility and increased aggregation in water, which competes with the target binding (ESI Table 3†). This manifests in low background uorescence, but also in
weak polymerized tubulin stabilization in vitro, low cytotoxicity
and poor staining. The delicate balance between spirolactonization–aggregation and target binding is close to ideal for 6-SiRCTX, the best performing SiR-based probe, which has probeD50 close to the dielectric constant value of water (ESI Table 3†).
We demonstrated the generality of the NGE cell permeability
enhancement by synthesizing probes targeting DNA and actin,
which can also operate at low nanomolar concentrations indicating unrestricted permeability, very high affinity and specicity of the new probes. This decreases the chances for unspecic probe accumulation in the membranes. Although invisible, such off-targeting might disturb physiological processes.
In conclusion, we employed the neighbouring group effect in the 40-regioisomer of rhodamine uorophores to create, to the best of our knowledge, the most efficient probes for tubulin, actin and DNA labelling in living cells (Table 1 and ESI Table 7†). Their excellent spectroscopic properties, outstanding cell permeability and specicity make them powerful tools for many types of microscopy techniques, including super-resolution methods. This simple approach of positional isomerism is applicable to rhodamine- and uorescein-type uorophores, does not require modication of the xanthene ring system and can be used for generating new probes and for improving the performance of existing ones.
Experimental
Measurements of absorbance spectra in 1,4-dioxane–water mixtures
Measurements of the absorbance changes in 1,4-dioxane–water
mixtures were performed by pipetting 2 mL stock solutions of
dyes or probes in DMSO into a 96 glass bottom well plate (11
wells per sample) made from propylene (Corning 3364). To the
wells going from right to le 300 mL of 1,4-dioxane–water
mixtures containing 100%, 90%, 80%, 70%, 60%, 50%, 40%,
30%, 20%, 10% or 0% 1,4-dioxane was added (if needed,
mixtures with 0.3% SDS are used, with an exception in 100%
dioxane due to solubility issues). Aer incubation for 1 hour at
room temperature, absorption of solutions in each well was
recorded from 320 nm to 850 nm with wavelength step size of
1 nm on a multiwell plate reader Spark® 20 M (Tecan). The
background absorption of the glass bottom plate was measured
in wells containing only the 1,4-dioxane–water mixture with
a similar amount of DMSO and subtracted from the spectra of
the samples. 3max versus dielectric constants (D) of 1,4-dioxane–
water mixtures were plotted.15 D50 value was obtained by tting
to dose–response equation EC50 as implemented in GraphPad
6.0 soware:
,
Hill!
A ¼ A0 þ ðAmax À A0Þ 1 þ D50
(1)
d
where A0 – absorbance at 3max at 3r ¼ 0, Amax – the highest reached absorbance at 3max, d – dielectric constant of 1,4dioxane–water mixture at a given point, Hill – Hill slope coefcient determining the steepness of a dose–response curve, D50 – corresponds to the d value that provokes half of the absorbance amplitude (Amax À A0).
Data points from the 4-TMR-LTX titration were tted to the bell-shaped dose–response curve described by the following
equation:
A ¼ A0 þ
Amax À A0
À
Amax À A0
(2)
D50 Hill1
A50 Hill2
1þ d
1þ d
7320 | Chem. Sci., 2020, 11, 7313–7323
This journal is © The Royal Society of Chemistry 2020
Open Access Article. Published on 22 June 2020. Downloaded on 7/10/2022 8:38:33 PM. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Edge Article
View Article Online
Chemical Science
where A0 – absorbance at 3max at 3r ¼ 0, Amax – the highest reached absorbance at 3max during the titration experiment. d – dielectric constant of 1,4-dioxane–water mixture at a given point, Hill1 – Hill slope coefficient determining the steepness of the ascending dose–response curve part, Hill2 – Hill slope coefficient determining the steepness of the declining dose– response curve part, D50 – corresponds to the d value that provokes half of the absorbance amplitude (Amax À A0) in the ascending dose–response curve part, A50 – corresponds to the d value that provokes half of the absorbance amplitude (Amax À A0) in the declining dose–response curve part.
Quantum chemical calculations
The initial geometries of the studied molecules were generated by using a molecular mechanics method (force eld MMF94, steepest descent algorithm) and a systematic conformational analysis as implemented in Avogadro 1.1.1 soware. The minimum energy conformer geometries found by molecular mechanics were further optimized with the Gaussian 09 program package28 by means of density functional theory (DFT) using the Becke, 3-parameter, Lee–Yang–Parr (B3LYP)29 exchange-correlation hybrid functional with 6-311G++(d,p) basis set,30 including the polarizable continuum model31 to simulate the environment in 1,4-dioxane or water. Furthermore, harmonic vibrational frequencies were calculated to verify the stability of the optimized geometries. All the calculated vibrational frequencies were real (positive) indicating the true minimum of the calculated total potential energy of the optimized system. The computations were performed at the High Performance Computing Center in G¨ottingen provided by GWDG.
Maintenance and preparation of cells
Human primary dermal broblasts were cultured in highglucose DMEM (Thermo Fisher, #31053044) and 10% FBS (Thermo Fisher, #10082147) with the presence of 1% penicillin– streptomycin (Sigma, #P0781) in a humidied 5% CO2 incubator at 37 C. HeLa cells were cultured in high-glucose DMEM (Thermo Fisher, #31966047) and 10% FBS (Bio&SELL #S0615), 1 mM sodium pyruvate (Sigma, #S8636) with the presence of 1% penicillin–streptomycin (Sigma, #P0781) in a humidied 5% CO2 incubator at 37 C. The cells were split every 3–4 days or at conuence. Cells were seeded in glass bottom 12-well plates (MatTek, #P12G-1.0-14-F) before imaging experiments. The U2OS cells were cultured in McCoys 5A medium (Thermo sher #16600082) and 10% FBS (Bio&SELL, #S0615) with the presence of 1 mM sodium pyruvate (Sigma, #S8636) and 1% of penicillin– streptomycin (Sigma #P0781) in a humidied 5% CO2 incubator at 37 C.
Cells were stained with the uorescent probes in DMEM (Thermo Fisher, #31053044) supplemented with 10% FBS ((Bio&SELL, #S0615)) at 37 C and 5% CO2. If needed, the cells were washed 2 times with HBSS (Lonza, #BE10-527F) and imaged in DMEM with 10% FBS. No-wash experiments were performed in DMEM with 10% FBS aer probe addition and incubation for the indicated period of time. For Airyscan
experiments, cells were seeded in a 10-well plate (Greiner bioone, #543079) 1 day before staining. The probes were applied to the cells in DMEM medium, and aer 1 h or overnight incubation cells were imaged without washing.
STED microscope with a 775 nm laser
Comparative confocal and STED images were acquired on an Abberior STED 775 QUAD scanning microscope (Abberior Instruments GmbH) equipped with 561 nm and 640 nm 40 MHz pulsed excitation lasers, a pulsed 775 nm 40 MHz STED laser, and an UPlanSApo 100Â/1.40 Oil objective. The following detection windows were used: for the TMR/580CP channel 615/20 nm, and for the 610CP/SiR channel 685/70 nm. In this setup, voxel size was 15–30 nm in the xy plane and 150 nm in the z-axis for 2D STED images. For the STED imaging with a 0.25 AU pinhole, the pixel size was set to 7 Â 7 nm and images were acquired with 2 line accumulation. Laser powers were optimized for each sample. 3D STED images were acquired using a pinhole set to 0.8 AU, voxel size set to 40 Â 40 Â 40 nm, 3D STED doughnut set to 90%, with single line accumulation and xzy scanning mode.
Estimation of the STED effect was performed by varying the STED laser power from 0 to 100% while measuring cells stained with tubulin probes. Obtained data were tted using GraphPad Prism 6 to the following equation:
Y ¼ pffiffiffidffiffifficffioffiffinffifffiffiffiffiffiffiffiffiffi
(3)
1 þ I =Isat
where dconf – confocal resolution, I – STED laser intensity power, Isat – saturating STED laser intensity power.
STED microscope with a 660 nm laser
STED images with a 660 nm STED laser were acquired on a Leica TCS SP8 STED 3X scanning microscope (Leica Microsystems) equipped with an HC PL APO CS2 100Â/1.40 oil objective and the incubator set to 37 C and 5% CO2. The 4-TMR-LTX probe was excited with the 561 nm, 80 MHz pulsed line of a white light laser, de-excited with a continuous wave 660 nm STED laser and detected in the 600/50 nm window. In this setup, the pixel size was set to 23 nm in the xy plane. The STED effect was estimated as described above.
Airyscan microscope
Live-cell imaging was performed on a Zeiss LSM880 system (Carl Zeiss) equipped with an oil immersion Plan-Apochromat 63Â/1.40 Oil Corr M27 objective and the incubator set to 37 C and 5% CO2. The following excitation lasers in combination with confocal or Airyscan acquisition regimes were used: argon laser 488 nm for 510R probes, 561 nm diode laser for TMR and 580CP probes, and 633 nm diode laser for 610CP and SiR probes. For TMR and 610CP, a combination of BP 570–620 + LP 645 lters with dichroic beam splitter SBS SP 615 in the TMR channel was used. The TMR, 510R and SiR combination was excited and scanned in sequence using BP 495–550 + LP 570 lters. Laser powers were optimized for each sample. Detailed excitation and detection schemes are shown in ESI Fig. 16.†
This journal is © The Royal Society of Chemistry 2020
Chem. Sci., 2020, 11, 7313–7323 | 7321
Categories
You my also like
Esi Scheme In Tamil Nadu
152.8 KB23K6.9KFundamentals of Green Chemistry: Efficiency in Reaction Design
3.5 MB1.5K595Advances in Understanding the Immune Imbalance betweenT
139.8 KB16.6K8KIn situ hybridisation: principles and applications
277.7 KB7.5K1.5KSupporting Information Photolytic Cleavage of Leader Peptides
1.9 MB8.6K1.6KSynthesis of novel pyrazolylbiscoumarin derivatives using
679.2 KB20K2KLone Pair Electrons Minimize the Lattice Thermal Conductivity
1.4 MB67.9K15.6KElectronic Supplementary Information revised
8.6 MB22.8K3KUnusual composition dependence of magnetic relaxation for Co
734.3 KB63.2K19.6K
