Ratiometric imaging of minor groove binders in mammalian cells using Raman microscopy

Quantitative drug imaging in live cells is a major challenge in drug discovery and development. Many drug screening techniques are performed in solution, and therefore do not consider the impact of the complex cellular environment in their result. As such, important features of drug–cell interactions may be overlooked. In this study, Raman microscopy is used as a powerful technique for semi-quantitative imaging of Strathclyde-minor groove binders (S-MGBs) in mammalian cells under biocompatible imaging conditions. Raman imaging determined the influence of the tail group of two novel minor groove binders (S-MGB-528 and S-MGB-529) in mammalian cell models. These novel S-MGBs contained alkyne moieties which enabled analysis in the cell-silent region of the Raman spectrum. The intracellular uptake concentration, distribution and mechanism were evaluated as a function of the pKa of the tail group, morpholine and amidine, for S-MGB-528 and S-MGB-529, respectively. Although S-MGB-529 had a higher binding affinity to the minor groove of DNA in solution-phase measurements, the Raman imaging data indicated that S-MGB-528 showed a greater degree of intracellular accumulation. Furthermore, using high resolution stimulated Raman scattering (SRS) microscopy, the initial localisation of S-MGB-528 was shown to be in the nucleus before accumulation in the lysosome, which was demonstrated using a multimodal imaging approach. This study highlights the potential of Raman spectroscopy for semi-quantitative drug imaging studies and highlights the importance of imaging techniques to investigate drug–cell interactions, to better inform the drug design process.


Materials and Methods. Thermal Melt Experimental
Salmon genomic DNA (gDNA; D1626, Sigma-Aldrich) at 1 mg/mL in 1 mM phosphate buffer (pH 7.4) containing 0.27 mM KCl and 13.7 mM NaCl (P4417, Sigma-Aldrich) was annealed at 90°C for 10 min and left to cool to room temperature. S-MGBs at 10 mM in DMSO were diluted with the same phosphate buffer to yield a single sample with 10 μM S-MGB and 0.02 mg/mL gDNA in 1 mM phosphate buffer containing 0.27 mM KCl and 13.7 mM NaCl. Control samples containing only S-MGB or gDNA were prepared, respectively. Samples were melted at a rate of 0.5°C/min from 45°C to 90°C with spectra recorded at 260 nm on a UV-1900 UVvis spectrophotometer fitted with a Peltier temperature controller (Shimadzhu) using LabSolutions (Tm Analysis) software. The melting temperatures (Tms) of the S-MGB:DNA complexes were determined by fitting a sigmoidal function using a Boltzmann distribution in OriginPro. Two independent experiments were carried out with values quoted with an error no worse than +1°C.

DNA Sample Preparation
DNA oligonucleotide sequence 5'-CGCATATATGCG-3' was purchased in lyophilized form from Alpha DNA, Canada) and used without further purification, purity assessed by NMR. 100 μM stock solutions of DNA were prepared with 150 mM ammonium acetate buffer solution (Fisher Scientific, Loughborough, Leicestershire, UK) and 2 mM potassium chloride solution (Fisher Scientific, Loughborough, Leicestershire, UK). This solution was annealed at 90 degrees for 10 minutes and allowed to cool to room temperature. 10 mM S-MGB stock in 100% DMSO (Sigma-Aldrich, St. Louis, MO, USA) were diluted to 1 mM S-MGB solution with 150 mM ammonium acetate. Final samples were prepared from this solution to yield final concentrations of 9 μM DNA, 100 μM KCl, and 100 μM s-MGB, 1% DMSO. DNA solutions containing no S-MGB included 1 % DMSO and were used as controls.

Mass Spectrometry Measurements
Native mass spectrometry experiments were carried out on a Synapt G2Si instrument (Waters, Manchester, UK) with a nanoelectrospray ionization source (nESI). Mass calibration was performed by a separate infusion of NaI cluster ions. Solutions were ionized from a thin-walled borosilicate glass capillary (i.d. 0.78 mm, o.d. 1.0 mm, (Sutter Instrument Co., Novato, CA, USA) pulled in-house to nESI tip with a Flaming/Brown micropipette puller (Sutter Instrument Co., Novato, CA, USA). A negative potential in range of 1.0 kV -1.2 kV was applied to the solution via a thin platinum wire (diameter 0.125 mm, Goodfellow, Huntingdon, UK). The following instrument parameters were used for the DNA: S-MGB-528 complex: capillary voltage 1.1 kV, sample cone voltage 90 V, source offset 110 V, source temperature 40 °C, trap collision energy 3.0 (V), trap gas 5 mL/min. For DNA: S-MGB-529 complex: capillary voltage 1.2 kV, sample cone voltage 90 V, source offset 110 V, source temperature 40 °C, trap collision energy 5.0 (V), trap gas 5 mL/min was used. For DNA with no MGB present, a capillary voltage 1.0 kV was applied to the sample. Sample cone voltage 80 V, source offset 95 V, source temperature 40 °C, trap collision energy 3.0 (V) and trap gas 4.0 mL/min was used. Data were processed using Masslynx V4.2 and OriginPro 2021, and figures were produced using chemdraw.

Raman spectroscopy
All Raman spectra were acquired on a Renishaw inVia Raman microscope equipped with a 532 nm Nd:YAG laser providing a maximum output of 50 mW and using a 1800 lines per mm grating, and a 785 nm diode laser providing a maximum output of 300 mW using a 1200 lines per mm grating. Prior to spectral acquisitions, the instrument was calibrated using the internal silicon standard at 520.5 cm −1 .

Investigating the alkyne Raman shift of S-MGBs in aqueous: organic mixtures
To investigate the responsiveness of S-MGB-528 and S-MGB-529 to the chemical composition, the alkyne Raman scattering of the compounds was investigated in aqueous:DMSO mixtures. S-MGBs were prepared at a concentration of 20 mM in water:DMSO (100:0 to 0:100 v/v). The solutions were transferred to a perfusion chamber system (Grace BioLabs) and Raman spectra were acquired using 532 nm excitation with a 60× lens (ca. 10 mW) for 0.5 s and 20 accumulations. Three replicate spectra were acquired per concentration.

Determining intracellular uptake concentration of S-MGBs
The alkyne Raman scattering of S-MGB-528 and S-MGB-529 in PBS solution, in the range of concentration between 0 and 52 mM, was measured. The measurements were performed using a perfusion chamber system (Grace BioLabs) under equivalent imaging conditions used for live cells. The measured linear relationship was then used to generate the equation used to estimate the in cellulo concentration of the compounds.

Intracellular uptake of S-MGBs using cellular permeabilisation
To investigate the role of the lipid membrane on the intracellular drug uptake, PNT2 and HeLa cells were seeded according to the Materials and Methods section, before washing with PBS (3×2 mL), fixation with 4% PFA in PBS (2 mL, 10 mins, 37°C) before further washing with PBS (2 mL x 3). The cells were either incubated with PBS in the presence or absence of TRITON-X-100 (0.1% v/v) for 10 mins before wasing with PBS (2 mL x 3). The cells were then incubated with DMEM or RPMI containing S-MGB-528 or S-MGB-529 (10 µM, 10 mins, room temperature) before washing with PBS (2 mL x 3) and imaging in PBS (4 mL).

Intracellular uptake in the presence of the metabolic inhibitor, NaN3
To investigate the presence of active transport, PNT2 and HeLa cells were treated with S-MGB-528 (10 µM, 4 h) and S-MGB-529 (20 µM, 4 h) in either RPMI or DMEM in the presence or absence of NaN3 (5 mM). Following the incubation, the cells were washed with PBS (3×2 mL) and imaged in PBS (4 mL).

Nuclear Magnetic Resonance (NMR) Experiments
All 1 H NMR and 13 C NMR spectral data for synthesised compounds were recorded using a Bruker DRX 500 spectrometer at 500 and 126 Hz respectively, with console advance III HD and 4.0.7 topspin software, using the deuterated solvent specified. Chemical shift values (δ) are expressed in parts per million (ppm). The following abbreviations are used for the multiplicity of the 1 H NMR signals: s (singlet), d (doublet), dd (doublet of doublets), dt (doublet of triplets), ddd doublet of doublets of doublets), m (multiplet), t (triplet), td (triplet of doublets), and q (quartet). Coupling constants are listed as J values, measured in Hz. Solvent references were DMSO-d6 referenced at 2.50 ( 1 H) and 39.52 ppm ( 13 C).

Infrared (IR) spectroscopy
IR spectra were run on a Shimadzu Corp. IRAffinity-1S Fourier Transform Infrared spectrophotometer fitted with a single reflection ATR accessory.

Freeze Dryer
Samples contained within an appropriately sized round bottomed flask were frozen by submerging in liquid nitrogen before being lyophilized using an LTE Mini Lyotrap Freeze Dryer.

Low-Resolution Mass Spectra (LR-MS)
Low-resolution mass spectra were recorded using an Agilent Technologies 1200 series LC-MS instrument with a 6130 single Quadropole and a dual electrospray and atmospheric chemical ionisation source. LC traces were recorded at a wavelength of 254 nm using an Agilent poroshell 120, EC-C18 2.7 μm, 4.6×100 mm column at 1 mL per minute. Method A was used for chromatographic separation, using solvent A (water with 5 mM ammonium acetate), and solvent B (MeCN with 5 mM ammonium acetate). Total lunch time: 18 min.

High-Resolution Mass Spec (HR-MS)
This was carried out at NMSF at Swansea University. Samples were solvated in MeCN, then diluted into either MeOH (salts) or MeOH + 30 mM ammonium acetate (neutrals) (to promote protonation and ammonium adduct formation rather than sodiation) for positive ion or diluted into either MeOH (salts) or MeOH + diethylamine (neutrals) for negative ion analysis. The Advion Triversa NanoMate (nano-electrospray) was used to deliver the appropriately diluted samples at a flow of approximately 0.25µL/min. This inlet is used with a 96-well plate, corresponding transfer tips and 400-nozzle spray-chip. Applied nanoelectrospray settings include: spray voltage 1.4 kV, gas pressure 0.4 psi, transfer capillary temperature 200 o C and transfer capillary voltage 30 V. Mass spectrometric detection was via a Thermo Scientific LTQ Orbitrap XL in positive/negative ionization modes. The instrument is externally calibrated each day using an in-house solution of Caffeine, MRFA (Met-Arg-Phe-Ala), and Ultramark 1621 for positive ion and this solution with added Sodium Dodecyl Sulfate and Sodium Taurocholate for negative ion. Internal calibration is accomplished via 'lockmass' of known background ions. Spectra were recorded at Resolution 100,000 (FWHM) over the m/z range 150 to 2000 Da with a mass accuracy of <3 ppm RMS.

Species m/z value Calculated mass of neutral species (Da) Single Stranded [SS]
3

Mass Spec Discussion
Native mass spectrometry experiments were used to demonstrate that both S-MGB-528 and S-MGB-529 bind to double stranded DNA, using a short DNA oligo with an AT rich binding site (5'-CGCATATATGCG-3'). nESI-MS was also used to confirm that both S-MGB-528 and S-MGB-529 bind to double stranded DNA as dimer species, as expected of molecules of this type.     Table S10: Estimated concentration of S-MGB-528 and S-MGB-529 in live PNT2 and HeLa cells with the relative enrichment ratio taking into account the extracellular drug concentration. Data represents mean ± SD. A minimum of three replicate images were acquired for each condition. Raman data were collected using 532 nm, 0.5 s, 10 mW, 60× immersive objective lens, step size 1 micron in x and y, 1 accumulation. This analysis readily resolves the nuclear (ratio 1 ≈ 0.2 au) and cytoplasmic regions (typically ratio 1 >0.25 au) of the fixed cells and the nuclear (ratio 1 ≤ 0.1 au) and cytoplasmic regions (typically ratio 1 > 0.1 au) of the fixed and permeabilised cells. Data represents mean ± SD (**** p ≤ 0.0001; *** p ≤ 0.001; ** p ≤ 0.01; * p ≤ 0.05; ns p > 0.05, Student's t test). A minimum of three replicate images were acquired for each condition. Scale bar size: 10 micron. Figure S8: Investigation of active transport as a potential mechanism of MGB uptake. (A) Representative Raman maps of PNT2 and HeLa cells treated with S-MGB-528 or S-MGB-529 in presence or absence of sodium azide. PNT2 and HeLa cells were alternatively preincubated and coincubated with medium containing sodium azide (5 mM, top row) or without sodium azide (bottom row), respectively. Raman scattering was collected using 532 nm, 0.5 s, 10 mW, 60× immersive objective lens, step size 1 micron in x and y, 1 accumulation. (B) Averaged Raman intensity of the alkyne group at 2104 cm -1 (S-MGB-528) and 2103 cm -1 (S-MGB-529) normalised to the symmetric C-H stretching (2930 cm -1 ) in live PNT2 and HeLa cells. PNT2 and HeLa cells were either preincubated and coincubated with or without sodium azide. Data represents mean ± SD (**** p ≤ 0.0001; *** p ≤ 0.001; ** p ≤ 0.01; * p ≤ 0.05; ns p > 0.05, Student's t test). A minimum of three replicate images were acquired for each condition. Figure S9: Investigation of the S-MGB uptake temperature dependence in live PNT2 cells.

Analysis
Representative Raman maps of live PNT2 cells after the treatment with S-MGB-528 and S-MGB-529 (10/20 µM, 4h, 4°C/26°C/37°C, in presence of HEPES or bicarbonate buffer). The chemical contrast is shown by the ratio 2850 cm -1 / (2850+2930) cm -1 that identifies the localisation of the cells and allows the investigation of the drug's intracellular distribution at sub-cellular resolution. The drug localisation is shown through the normalised alkyne Raman intensity (i.e. 2104 cm -1 /2930 cm -1 ). Raman data were collected using 532 nm, 0.5 s, 10 mW, 60× immersive objective lens, step size 1 micron in x and y, 1 accumulation.

Figure S10
: Alkyne scattering of S-MGB-528 and S-MGB-529 in live PNT2 cells at the cellular (top line) and sub-cellular resolution (bottom line). PNT2 cells were incubated with growing concentration of either S-MGB-528 or S-MGB-529 in a humified incubator (4 h). Raman data were collected using 532 nm, 0.5 s, 10 mW, 60× immersive objective lens, step size 1 micron in x and y, 1 accumulation. The alkyne Raman intensity was normalised to the symmetric C-H stretching (2930 cm -1 ). A minimum of three replicate images were acquired for each condition. The chemical contrast is shown by the ratio 2850 cm -1 / (2850+2930) cm -1 that identifies the localisation of the cells and allows the investigation of the drug's intracellular distribution at sub-cellular resolution. The drug localisation is shown through the normalised alkyne Raman intensity (i.e. 2104 cm -1 /2930 cm -1 ) to the symmetric C-H stretching (2930 cm -1 ). (C) Normalised alkyne intensity of S-MGB-528 and S-MGB-529 in HeLa and PNT2 cells in the presence or absence of the structural analogues. The Raman data were acquired using 532 nm, 0.5 s, 10 mW, 60× objective immersive lens, step size 1 micron in x and y, 1 accumulation. Data represents mean ± SD (**** p ≤ 0.0001; *** p ≤ 0.001; ** p ≤ 0.01; * p ≤ 0.05; ns p > 0.05, Student's t test). A minimum of three replicate images were acquired for each condition.