Spectroscopic Probing of Dynamic Changes during Stimulation and

May 19, 2007 - For example, short-term, dynamic changes in Raman bands during the ... For a more comprehensive list of citations to this article, user...
0 downloads 0 Views 316KB Size
Download PDF
Anal. Chem. 2007, 79, 4581-4587

Spectroscopic Probing of Dynamic Changes during Stimulation and Cell Remodeling in the Single Cardiac Myocyte O. Inya-Agha,† N. Klauke,*,† T. Davies,‡ G. Smith,‡ and J. M. Cooper†

Department of Electronics, University of Glasgow, Glasgow G12 8LT, Scotland, and Institute of Biomedical and Life Sciences, University of Glasgow, Glasgow G12 8QQ, Scotland

Optical microscopy, involving both fluorescence imaging and confocal Raman microspectroscopy, was used to visualize single, isolated, electrically active heart muscle cells. For example, short-term, dynamic changes in Raman bands during the contraction cycle, as well as persistent band changes during structural remodeling (microscopic rearragements of cellular structures) in culture over longer periods of time, were obtained from the cellular content (sarcoplasm) of the heart cells. The results of the short-term studies, collected during electrical stimulation, showed dynamic changes in the Raman amide I band intensity, which occurred in phase with changes in cell length during cardiomyocyte contraction. The longer term studies of quiescent cardiomyocytes in culture over 3 days revealed a progressive and sustained increase in the intensity of the amide I band. Over the same period of culture, a decrease in the number of t-tubules (invaginations of the cell membrane, sarcolemma, which ensure the spreading of the action potential into the bulk of the sarcoplasm) was observed using confocal z-sections of the fluorescently labeled sarcolemma. The ability to measure both short-term dynamic changes associated with stimulated contraction and longer term persistent remodeling in the structure of intracellular macromolecules is valuable for assessing the physiological state of the cell, in real time. Advanced optical microscopy, particularly when used in combination with specific fluorescent probes, allows for the investigation of biochemical reactions in biological systems, providing high temporal and spatial resolution information from activated single cells.1,2 By contrast, nonresonant Raman spectroscopy is considerably less invasive, as it does not require the use of specific probes and provides the potential to monitor the numerous small metabolites or chemical signatures in a living * To whom correspondence should be addressed. E-mail: [email protected]. † Department of Electronics. ‡ Institute of Biomedical and Life Sciences. (1) Xie, X. S.; Yu, J.; Yuan Yang, W. Science 2006, 312, 228-230. (2) Navratil, M.; Mabbott, G. A.; Arriaga, E. A. Anal. Chem. 2006, 78, 40054019. 10.1021/ac0622476 CCC: $37.00 Published on Web 05/19/2007

© 2007 American Chemical Society

single cell, which would otherwise be difficult to label without altering their cellular behavior.3 The Raman signal arises from the inelastic scattering of photons when they interact with matter, resulting in spectral responses that are shifted to wavelengths both longer and shorter than the wavelength of excitation. The extent of any measured shift depends on both the molecular structure and vibrational transitions of the scattering molecule, making the method useful for investigating conformational energetics. As a vibrational spectroscopic technique, Raman spectroscopy has a capacity for identification that is similar to IR, enabling the presence or absence of particular molecules to be assigned to known vibrational modes thereby providing a method for fingerprinting metabolic changes or for disease diagnosis,3 for example. Unlike IR absorption, however, Raman scattering is linear in intensity (and thus easier to quantify). It does not suffer from broad water interference and can therefore be used in aqueous environments, making it particularly well suited to monitoring living cells. However, the Raman effect is very weak as typically only 1 in 106-108 photons undergo an inelastic light scattering event. In addition, Raman spectra collected with a visible laser from biological samples can often be masked by autofluorescence which, while much broader than the sharp Raman bands, can dominate the spectrum and obscure the Raman signal. Such autofluorescence often therefore needs to be removed by signal processing (background substraction). When monitoring proteins in situ, changes in protein secondary structure involving intramolecular rearrangements will affect the bond strengths and bond angles within the protein and, as such, will appear as wavenumber shifts of the Raman bands. As a consequence, Raman spectroscopy is already well established as a method for the determination of protein secondary conformation, e.g., β-sheet or R-helix secondary structures.4-10 In addition, (3) Ellis, D. I.; Goodacre, R. Analyst 2006, 131, 875-885. (4) Lippert, J. L.; Tyminski, D.; Desmeules, P. J. J. Am. Chem. Soc. 1975, 98, 7075-7080. (5) Gaber, B. P.; Peticolas, W. L. Biochim. Biophys. Acta 1977, 465, 260-274. (6) Harada, I.; Takeuchi, H. In Advances in Spectroscopy; Clarke R. J. H., Hester, R. E., Eds,; Wiley: New York, 1986; pp 113-175. (7) Miura, T.; Thomas, G. J. In Subcellular Biochemistry; Biswas B. B., Roy, S., Eds,; Plenum Press: New York, 1995; pp 55-99. (8) Overman, S. A.; Thomas, G. J. J. Raman Spectrosc. 1998, 29, 23-29. (9) Uzunbajakava, N.; Lenferink, A.; Kraan, Y.; Willekens, B.; Vrensen, G.; Greve, J.; Otto, C. Biopolymers 2003, 72, 1-9.

Analytical Chemistry, Vol. 79, No. 12, June 15, 2007 4581

polarized Raman spectroscopy can measure the tilt angle of a specific vibrational mode and thus provide information about the molecular orientation distribution of functional groups in polypeptides.11 In confocal Raman microspectroscopy, the volume of analysis is represented by the in-focus laser, making it an ideal tool for the mapping of single cells.10,12-15 However, determining the temporal and spatial changes of the intracellular distribution of unlabeled molecules has proved difficult due to the weak signal of the Raman scattering. To compensate for the low scattering cross sections, resonance enhancement, surface enhancement, and coherent anti-Stokes Raman scattering (CARS) have also been applied to microspectroscopy. However, resonance Raman spectroscopic probes are limited in use to enhancing the detection of selected molecules within cells, while CARS, which requires two picosecond pulsed laser lines for coherent signal generation,16 is neither a routine nor a widely available method. Resonant Raman imaging has also been performed by scanning the sample with a fine metal tip and recording the field-enhanced Raman scatter with a near-field optical microscope.17,18 This technique, which relies upon surface enhancements, probes the dynamics of structural changes on the cell surface, e.g., extracellular, which occur within the evanescent field of the tip. Intracellular measurements of surface-enhanced Raman scattering have been achieved by introducing functionalized silver nanoparticles into the cell,19 although these have proved difficult to target into selected subcellular compartments and indeed may change the function and longevity of the cell. As stated, and in contrast to enhanced methods, normal Raman spectroscopy can be generally applied to the identification of many intracellular molecules using commercially available spectrometers and minimal sample preparation. Indeed, it is the recent dramatic advances in the sensitivity of CCD detectors that has enabled the technique to become useful as a noninvasive method for both cell biology and medical diagnosis, allowing samples either to be maintained in close-to-physiological conditions or to be used in vivo.20-22 The study of isolated adult ventricular myocytes has been central to the understanding of the physiology and pathophysiology of the heart muscle. We now show that nonresonant confocal Raman microspectroscopy can probe for changes in intracellular (10) Short, K. W.; Carpenter, S.; Freyer, J. P.; Mourant, J. R. Biophys. J. 2005, 88, 4274-4288. (11) Rousseau, M.-E.; Beaulieu, L.; Lefevre, T.; Paradis, J.; Asakura, T.; Pezolet, M. Biomacromolecules 2006, 7, 2512-2521. (12) Schuster, K.; Reese, I.; Urlaub, E.; Gapes, J.; Lendl, B. Anal. Chem. 2000, 72, 5529-5534. (13) Arzhantsev, S. Y.; Chikishev, A. Y.; Koroteev, N. I.; Greve, J.; Otto, C.; Sijtsema, N. M. J. Raman Spectrosc. 1999, 30, 205-208. (14) Freeman, T. L.; Cope, S. E.; Stringer, M. R.; CruseSawyer, J. E.; Batchelder, D. N.; Brown, S. B. J. Raman Spectrosc. 1997, 28, 641-643. (15) Krafft, C.; Knetschke, T.; Siegner, A.; Funk, R.; Salzer, R, Vib. Spectrosc. 2003, 32, 75-83. (16) Nan, X.; Potma, E. O.; Xie, X. S. Biophys. J. 2006, 91, 728-735. (17) Bouhelier, A. Microsc. Res. Tech. 2006, 69, 563-579. (18) Neugebauer, U.; Roesch, P.; Schmitt, M.; Popp, J.; Julien, C.; Rasmussen, A.; Budich, C.; Deckert, V. Chem. Phys. Chem. 2006, 7, 1428-1430. (19) Talley, C. E.; Jusinski, L.; Hollars, C. W.; Lane, S. M.; Huser, T. Anal. Chem. 2004, 76, 7064-7068. (20) Shim, M. G.; Song, L.; Marcon, N. E.; Wilson, B. C. Photochem. Photobiol. 2000, 72, 146-150. (21) Puppels, G. J. Microbeam Anal. 2000, 165, 63-64. (22) Caspers, P. J.; Lucassen, G. W.; Puppels, G. J. Biophys. J. 2003, 85, 572580.

4582

Analytical Chemistry, Vol. 79, No. 12, June 15, 2007

myocyte structure under near-physiological conditions, e.g., during the contraction cycle. We also demonstrate that the method can be used to probe cell remodeling by relating the Raman spectra to standard physiological markers for myocyte function, including cell-shortening and sarcolemmal integrity. The use of the confocal Raman configuration allowed us to select regions of the cell and, specifically, to exclude the nuclei (with its very strong Raman signatures from DNA, thereby reducing the complexity of the Raman signal). It has been known for some time that the isolation and culture of cardiac myocytes leads to profound changes in cell structure and function.23 For example, a major cellular compartment, the cell membrane (sarcolemma) with its regular invaginations (ttubules) is affected during culture such that the invaginations detach from the surface membrane. This results in the asynchronous excitation-contraction coupling due to the lack of a uniform calcium influx into the bulk of the sarcoplasm during the propagating action potential.24,25 This phenomenon is one example characteristic of cellular remodeling, occurring not only in cell culture but also in situ (e.g., the diseased heart). Understanding this process is therefore a significant challenge for studying primary cells in conditions that are as close to physiological as possible. The use of confocal Raman microspectroscopy to investigate the sarcoplasm during long-term culture to probe changes in myocyte function and structure during dedifferentiation (remodeling) in cell culture will prove of value, in this latter respect. MATERIALS AND METHODS Cell Isolation. Adult rabbit cardiomyocytes were obtained by collagenase digestion in combination with mechanical agitation, as described previously.26 Following cell isolation, calcium-tolerant cardiomyocytes were placed in M199 serum-free medium with Earle’s salts at a concentration of 50 000 rod-shaped cells mL-1. The cells were seeded into transparent slide flasks (Ibidi GmbH, Munich, Germany) coated with 20 mg mL-1 laminin (Sigma Aldrich, Dorset, UK) to facilitate cell adhesion. Cell Stimulation. Cardiomyocytes, which are normally quiescent in culture, were induced to contract synchronously via electrical field stimulation as described previously.27 The electrodes were high-purity stainless steel needles (2-mm diameter) submersed 5 cm apart at opposite ends of the slide flask. The polarity of the electrodes was alternated with each electrical pulse to minimize polarization. In order to set the threshold voltage for contraction, the lowest voltage required to stimulate greater than 50% of the cells was determined and then exceeded by 50%, resulting in an approximate response ratio of 80%. The resultant pulses were ∼6 V cm-1 in amplitude, 5-ms duration at 0.5 Hz. Cardiomyocytes used for these optical studies were then selected on the basis of several additional criteriasnamely, that they responded faithfully to the low-amplitude field stimulus (5 V cm-1) with regular and uniform shortening, that they had regular and (23) Volz, A.; Piper, H. M.; Siegmund, B.; Schwartz, P. J. Mol. Cell. Cardiol. 1991, 23, 161-173. (24) Brette, F.; Orchard, C. Circ. Res. 2003, 92, 1182-1192. (25) Leach, R. N.; Desai, J. C.; Orchard, C. H. Cell Calcium 2005, 38, 515-526. (26) Mitcheson, J. S.; Hancox, J. C.; Levi, A. J. Cardiovasc. Res. 1998, 39, 280300. (27) Berger, H. J.; Prasad, S. K.; Davidoff, A. J.; Pimental, D.; Ellingsen, O.; Marsh, J. D.; Smith, T. W.; Kelly, R. A. Am. J. Physiol. 1994, 266, H341-H349.

clearly defined striation patterns with no obvious signs of damage in the intercalated disk regions, and that they were 20-30 µm wide and 140-180 µm long. The duration of the contractionrelaxation cycle was ∼900 ms. Cell Length Measurement. Cell lengths were monitored using an Ionoptix (Milton, MA) image analysis system as previously described.28 Quiescent Long-Term culture. Cells were maintained in medium 199 (Sigma-Aldrich, Dorset, UK) at 37 °C in a saturated atmosphere comprising air supplemented with 5% CO2. The culture medium was changed every 24 h, and then immediately prior to cell analysis, to ensure removal of any nonadherent cells and of all cell debris or extracellular matrix. Measurement of t-Tubular Density Using di-8-ANEPPS Staining. Isolated cardiomyocytes were suspended for 10 min in Tyrodes solution containing 10 µmol L-1 di-8-ANEPPS (pyridinium, 4-[2-[6-(dioctylamino)-2-naphthalenyl]ethenyl]-1-(3-sulfopropyl)-, inner salt, Invitrogen, Paisley, UK), a potential-sensitive aminonaphthylethenylpyridinium dye. After washing, the cells were placed on a coverslip bottomed bath (200-µL volume) of a Nikon Eclipse inverted microscope using a (Nikon Inc., Melville, NY) Nikon 40× oil objective (NA 1.4). Confocal images of di-8-ANEPPS fluorescence were obtained with a BioRad Radiance 2000 (BioRad Inc, Herts, UK) using the 488-nm line of an Ar-Kr laser. The iris diameter was set at 1.9 providing an axial (z) resolution of ∼0.9 µm and X-Y resolution of ∼0.5 µm based on full width halfmaximal amplitude measurements of images of 0.1-µm fluorescent beads (Molecular Probes Europe, Leiden, Netherlands). Fluorescence image data were acquired in image-scan mode at 2 ms/line, 512 pixels/scan, 512 lines/image. Each section was Kalman averaged 4 times. A series of images in the z-axis was taken every 0.5 µm from the top surface of the cardiomyocyte to the bottom. Pixels associated with peripheral sarcolemma (i.e., around the edge of the cell, avoiding areas of intercalated disk) were sampled, and the intensity distribution of these pixels was used to determine the threshold for pixels within the cell that could be attributed to the membrane. Pixels with an intensity greater than or equal to the mean pixel value of the peripheral sarcolemma were attributed to the t-tubule. This criterion allowed the generation of a binary image of the cell with white pixels presenting areas of internal membrane. Based on the original cell boundaries, the longitudinal axis of the cell was split into 10 equal segments and the percentage of pixels positive for di-8-ANEPPS staining was calculated relative to the total number of pixels occupied by the width of the cell in that segment. This analysis was applied to the central 10-15 sections (5-7.5-µm slice) in each cardiomyocyte, thus avoiding oblique sections on the upper and lower surfaces and ensuring that measurements were made from a region where cell length was uniform. Raman Spectroscopy. Raman spectra were acquired using a LabRAM 300 confocal Raman microscope system (Jobin Yvon, Middlesex, UK). The sample interface was a fully integrated Olympus microscope equipped with Olympus objectives (Olympus UK Ltd., London, UK), coupled to a 300-mm-focal length spectrometer. The spectrometer was equipped with a grating providing a resolution of 3.5 cm-1/pixel. An integrated He/Ne laser operating at 17 mW provided 633-nm excitation. The instrument

was calibrated at the factory according to the manufacturer’s standard procedure. Wavenumber precision and spectral resolution was rechecked in the laboratory before each experiment by measuring the full width at half-maximum of the 520.0-cm-1 band of crystalline silicon.29 A 100×/0.8 NA objective was used to obtain photomicrographs, and a 50×/0.55 NA objective was used with a confocal aperture of 150 µm to provide a confocal volume of ∼2 µm3 for spectral acquisition. The detector was an open electrode, thermoelectrically cooled charge-coupled device detector (Andor Technology, Belfast, Northern Ireland). Spectra were acquired from electrically stimulated cardiomyocytes over 50 s in two integration time sets. The first set was acquired with a 1-s integration time to include both contracted and relaxed states within a single spectrum (Figure 2); the second set (Figure 3) was acquired with 300-ms integration time. Spectra of cells in culture were acquired at 3 × 2 s integration times to maximize S/N ratio (Figure 5). In all cases, spectra were acquired from the cell cytoplasm close to the middle of the cell avoiding detection of nuclei and perinuclear regions (Figure 1a).15 Raman spectral data were analyzed using Matlab (The Mathworks Inc., Natick, MA) using vendor-supplied routines. All spectra were baseline corrected and normalized to the intensity of the 1447-cm-1 band, which was used as an internal intensity standard because of its insensitivity to conformational changes.4

(28) Klauke, N.; Smith, Cooper, G. L. J. Biophys. J. 2003, 85, 1766-1774.

(29) Beeman, D.; Tsu, R.; Thorpe, M. F. Phys. Rev. B 1985, 32, 874-878.

Figure 1. (a) Photomicrograph of a single isolated adult cardiomyocyte. The circle indicates an example of the region from which spectra were acquired, although the two nuclei of the cardiomyocyte were avoided during spectral acquisition. (b) Raw, unmodified Raman spectra acquired from single cardiomyocytes using different integration times, with 300 ms in (i), 1 s in (ii). Spectrum iii is acquired with 2-s integration from a cell in day 2 of long-term culture. Spectrum iv is an averaged, baseline-subtracted spectrum of 7 cardiomyocytes from day 0, before any electrical stimulation (50-s integration/cell).

Analytical Chemistry, Vol. 79, No. 12, June 15, 2007

4583

Figure 2. (a) Plot of cell length against time, showing cell shortening events resulting from a standard train of 25 electrical stimuli applied to induce cell shortening. All sample regions represent equal times of 50 s. (b) Band areas of three vibrational bands (∼922, ∼1447, and ∼1658 cm-1) during the three sample periods showing the band area changes due to cell shortening. The amide I band is the only one that shows significant change in area.

Additional data manipulation included standard signal averaging, background subtraction, and cosmic event (spike) removal. RESULTS AND DISCUSSION Raman Spectrum of an Isolated Quiescent Cardiomyocyte. A photomicrograph of a single isolated adult cardiomyocyte is shown in Figure 1a. The circle indicates a representative area from which spectra were acquired. Parts b and c in Figure 1 serve to illustrate both the nature of the spectra under different conditions and the changes in signal/noise of spectra collected with different integration times. For example, trace i in Figure 1b and traces i and ii in Figure 1c are raw, unmodified spectra acquired from a single cardiac myocyte using different integration times of 300 ms (Figure 1b,i), 1 s (Figure 1c,i), and 2 s (Figure 1c,ii), respectively. In Figure 1c, spectrum i was acquired on day 0, and spectrum ii was acquired on day 2 of the long-term culture. In Figure 1b, spectrum ii is baseline-subtracted and averaged over seven cardiomyocytes (50-s acquisition per cell) from day 0, before either quiescent culture or electrical stimulation. This latter spectrum also shows the Raman band assignments, which are presented in Table 1.4,6-8,13-15,21,30-35 (30) Shim, M. G.; Wilson, B. C. J. Raman Spectrosc. 1997, 28, 131-142. (31) Caspers, P. J.; Lucassen, G. W.; Wolthuis, R.; Bruining, H. A.; Puppels, G. J. Biospectroscopy 1998, 4, S31-S39. (32) Wood, B.; McNaughton, D. J. Raman Spectrosc. 2002, 33, 517-523. (33) Carew, E. B.; Stanley, H. E.; Seidel, J. C.; Gergely, J. Biophys. J. 1983, 44, 219-224. (34) Buschman, H. P.; Deinum, G.; Motz, J. T.; Fitzmaurice, M.; Kramer, J. R.; van der Laarse, A.; Bruschke, A. V.; Feld, M. S. Cardiovasc. Pathol. 2001, 10, 69-82.

4584 Analytical Chemistry, Vol. 79, No. 12, June 15, 2007

Figure 3. (a) Cell length during a single contraction cycle, plotted in (i). (ii) shows the spectra obtained by gating the acquisition into 300-ms time periods relative to the cardiomyocyte twitching. Each spectrum is matched with the length of the cardiomyocyte at the time of acquisition. (b) Cell length and amide I band area data are presented in a graphical format. Amide I band area is compared to internal standard (∼1447-cm-1 band) area.

The bands appearing at ∼827 and ∼855 cm-1 are due to tyrosyl residues in the cardiomyocyte proteins. The ratio of their intensities (I855/I827) can be used to determine the average oxidation state of tyrosine residues in a protein.7 The ∼886-cm-1 band was assigned to a tryptophan normal mode involving both a deformation of the six-membered ring and a displacement of the NH group along the direction of the N-H bond.6 The ∼922-cm-1 band was also assigned to tryptophan.36 The ∼1003-cm-1 phenylalanine band has been previously reported as a marker of cell death.37 The (35) Wang, Y.; Purrello, R.; Jordan, T.; Spiro, T. G. J. Am. Chem. Soc. 1991, 113, 6359-6368. (36) Dou, X.; Yamaguchi, Y.; Yamamoto, H.; Doi, S.; Ozaki, Y. J. Raman Spectrosc. 1998, 29, 739-742. (37) Verrier, S.; Notingher, I.; Polak, J. M.; Hench, L. L. Biopolymers 2004, 74, 157-162.

Figure 4. (a) Transmission image of typical single cardiomyocytes at day 0, day 1, and day 2 of quiescent culture. (b) Corresponding confocal images from these cells illustrating the staining of the t-tubules by the membrane-bound dye di-8-ANEPPS. (c) Binary conversions of the confocal images that are used to assess t-tubule density showing a clear decrease in t-tubules density over 48 h of culture. (d) Mean values of cell area stained with Di-8-ANEPPS. (e) Cell volume is shown to remain constant over the duration of the experiment.

∼1448-cm-1 methylene bending mode has been shown to be insensitive to the conformation of proteins38,39 and was used as an internal standard to quantify the intensities of other bands. The ∼1480-cm-1 shoulder on this band is assigned to tryptophan.36 Disordered structure amide I is expected in the ∼1655-1665-cm-1 interval, with R-helix amide I expected between ∼1645 and 1655 cm-1.7 The cardiomyocyte amide I band observed at ∼1658 cm-1 in this study indicates that disordered structures are the main conformation of proteins in the sarcoplasm. Changes in the Cardiomyocyte Spectrum on Field Stimulation. As stated, all cell stimulation experiments were performed within 4 h after cell isolation, during which culture period no spontaneous band shifts were observed. Cells were electrically stimulated with a train of stimuli, each stimulus causing a transient shortening as shown in Figure 2a. Figure 2b plots examples of three vibrational band areas (∼922, ∼1448, and ∼1658 cm-1) to the three sample periods, e.g., before (1), during (2) and after (3) stimulated shortening. Samples 1 and 3 are before and after stimuli, respectively, while sample 2 was acquired during the stimulus period, over which time the cell shortened 25 times (at a frequency of 0.5 Hz). All sample regions represent the summation of spectra over equal times of 50 s. The ∼1448-cm-1 methylene band, which was the internal standard, is shown in all instances, as a reference. Observation of the change in the amide I band, ∼1658 cm-1, acquired over these periods, led us to collect sets of fast (300 ms) spectra for more detailed investigation (this was the minimum (38) Barrett, T. W.; Peticolas, W. L.; Robson, R. C. Biophys. J. 1978, 23, 349358. (39) Frushour, B. G.; Koenig, J. L. Biopolymers 1974, 13, 783-813.

acquisition time sufficient to give acceptable signal-to-noise ratios). Figure 3a (ii) shows six such spectra which were sequentially acquired at 300-ms intervals during the 2-s contraction cycle shown in Figure 3a (i). These data are presented in a graphical format in Figure 3b, in which the changes in the amide I region are correlated to the cardiomyocyte shortening (Figure 3b, top), showing that there were pronounced changes in the amide I band area (Figure 3b, middle), accompanying cell shortening, while the signal from the internal standard (∼1448 cm-1) remains relatively constant (Figure 3b, bottom) These results show that the temporal resolution of the confocal Raman microscope sufficient for the real-time probing of the densely packed contractile proteins within the mycoyte with good signal-to-noise ratio, in spite of the short duration of the myocyte shortening. Studying invertebrate muscle, Caille et al. found that cell contraction increased the scattering intensity of their ∼1650-cm-1 amide I bands, as well as of the ∼939-cm-1 C-C or C-N stretch.40 Both observations were attributed either to an increase of the R-helical content of the conformation of the contractile proteins or to a partial reorientation of the R-helical segments of the contractile proteins during contraction. Such a reorientation would result in the amide I carbonyl groups having a parallel electric field alignment with that of the incident radiation, causing increased intensities. The authors further reported a diminished scattering intensity when the incident electric field was aligned perpendicularly to muscle fibers.40 Based on this, the observed decreases in amide I scattering intensity during the cardiomyocyte shortening in this study was attributed to decreasing alignment (40) Caille, J. P.; Pigeongosselin, M.; Pezolet, M. Biochim. Biophys. Acta 1983, 758, 121-127.

Analytical Chemistry, Vol. 79, No. 12, June 15, 2007

4585

Table 1. Band Position (cm-1) Assignment and Vibrational Mode 827 855 886 922 939 1003 1010 1030 1037 1042 1111 1125 1154 1167 1210 1230-1310 1299 1339 1448 1480 1587 1607 1658 a

Figure 5. (a) Shows changes in Raman spectra of cardiomyocytes over 3 days. Most notable, the Trp band shifts and the Phe and amide I band increases. (b) Plots of the areas of the ∼922-, ∼1003-, and ∼1658-cm-1 bands, as a function of time, showing that while the ∼922-cm-1 tryptophan band remains constant, both the phenylalanine and amide I bands increase with time.

between the disordered structures in the sarcoplasm, occurring as a consequence of protein rearrangements. The reduced degree of disorder during cell shortening may cause more perfect misalignment between the excitation laser and the contractile protein filaments and a concomitant reduction in amide I band intensity. Changes in t-Tubule Density of Myocytes in Long-Term Quiescent Culture. Culture of primary myocytes is associated with significant changes in cardiomyocyte ultrastructure and function, collectively termed cellular remodeling. An example of this is the rounding at the ends of cells observed in transmission light microscopy during this study (Figure 4a). Confocal fluorescence imaging of the cardiomyocyte membrane (sarcolemma) was performed to follow the time course of another marker for cell remodeling called detubulation, e.g., the detachment of the membrane invaginations (t-tubules) from the surface membrane (Figure 4b). In unimpaired cardiomyocytes, the synchronous propagation of the depolarization during an action potential from 4586 Analytical Chemistry, Vol. 79, No. 12, June 15, 2007

tyrosinea tyrosinea tryptophan ring deformation and NH displacement tryptophan CsC skeletal (R-helix backbone) stretch phenylalanine tryptophan phenyl CH bend CsC or C-H bending C-N stretch tryptophan pyridine half-ring asymmetry CsC stretch of isoleucine tyrosine combined phenylalanine and tyrosine amide III region NH bend and CH stretch CH2-CH2 stretch and CH2-CH3 bend tryptophan protein conformation insensitive CH2-CH2 bend tryptophan amide II CN stretch tyrosine amide I

Fermi doublet.

the cell surface into the bulk of the sarcoplasm is mediated by the t-tubules, a network formed inside the cardiomyocyte through the invaginations of the sarcolemma. A similar loss of t-tubules has also been noted by others, but over much longer time periods.41 Representative transmission micrographs of individual cardiomyocytes at day 0, day 1, and day 2 of the quiescent cell culture are shown in Figure 4a (i-iii). The median optical sections from the same cells illustrate the staining of the t-tubules by the membrane-bound di-8-ANEPPS in Figure 4b (i-iii). In Figure 4c, the binary conversions, e.g., conversions from gray scale to black and white according to a given threshold, of the confocal images were used to quantify the t-tubule density and show clearly that the number of t-tubules decreases over 48 h of culture. The mean values of cell area stained with Di-8-ANEPPS are shown in Figure 4d, while Figure 4e shows that the cell volume remained constant during the course of the experiment. With progressive t-tubule loss, the protein arrangement in the cardiomyocyte appears to acquire more disordered structure, as indicated by the increase in the intensity of the ∼1658-cm-1 band (Figure 5). Direct comparison between the fluorescence and Raman data sets underlines the fact that the Raman spectra changed over the same time course as the t-tubule density. Future work will correlate more specific changes of the Raman spectra with the fluorescently identified changes of cellular structure or metabolism as an unambiguous marker of cell remodeling during long-term culture. Changes in Raman Spectra on Long-Term Quiescent Culture. As evidenced from Figure 5, the largest Raman spectral changes that occurred between day 0 and day 1 are the tryptophan band shifts, along with a reduction in the intensities of the ∼886and ∼922-cm-1 bands and the increase in the phenylalanine band (41) Mitcheson, J. S.; Hancox, J. C.; Levi, A. J. Pflugers Arch. 1996, 431, 814827.

intensity. Another notable difference was the appearance of the Fermi doublet at 827/855 cm-1. The amide I band area remained relatively constant until day 3 when it showed a sharp increase in intensity. These and other changes to Raman spectra are further discussed below. The evidence for cellular remodeling apparent from the irreversible changes in the different regions of the Raman spectra is detailed below. For example, the tyrosyl Fermi doublet appearing at ∼855 and 827 cm-1 was more prominent from day 1 onward. The I855/I827 ratio for this doublet was found to range from 0.35 to 0.45, indicating a tendency of the phenolic OH to change from donor status toward acceptor status. This is believed to correlate to changing nearest-neighbor relationships as protein side chains move around with respect to each other during conformational rearrangements within the cell. Residues can orient themselves next to other residues that are of different electronegativity than previous neighbors or may re-orient differently toward the same neighbors, so that a change in electron acceptance/donation relationships is necessary to maintain thermodynamic stability of the new conformation. In Figure 5a, the ∼886- and ∼922-cm-1 bands of tryptophan decrease over time, while the ∼1003- (Phe) and the ∼1658-cm-1 (amide I) bands increased in band area. The C-C/C-N band observed at ∼939 cm-1 is a shoulder on the much larger tryptophan band measured at ∼922 cm-1 on day 0. After the tryptophan band shift occurs, the ∼939-cm-1 band appeared smaller. Areas were difficult to plot for this band due to its low signal-to-noise ratio. In Figure 5b, the same data were therefore presented in a bar graph for clarity. Tryptophan bands have a strong influence on protein spectra even when tryptophan was only present in low proportion in comparison to other residues. This is due to the conjugated group in the molecule, which is subject to strong change in polarizability, resulting in a stronger presence in the protein Raman spectrum than other, less-electronrich, moieties. Band shifts were observed between days 0 and 1 in the ∼886- (tryptophan, Trp), ∼922- (Trp), and ∼1480-cm-1 (Trp) bands. The ∼886-cm-1 band’s position is sensitive to N-H hydrogen bond donation within the range from ∼886 cm-1, which indicates no hydrogen bonding, to ∼871 cm-1, which indicates strong hydrogen bonding.42 Long-term quiescent culture increases the average hydrogenbonded state of cardiomyocyte tyrosines, shown by a band shift from ∼883 cm-1 on day 0 to ∼886 cm-1 on day 1. The hydrogen bonds, once formed, are not reversed, as evidenced by the fact that the ∼886-cm-1 band experiences no further changes for the duration of this study. The ∼928-cm-1 band shifts to ∼922 cm-1 and the ∼1480-cm-1 band shows up as a shoulder on the stronger methylene band measured at ∼1448 cm-1. The ∼939-cm-1 band (C-C/C-N) is notable in this study because it remains as an unshifted frame of reference for the behavior of its nearest neighbor, the tyrosine band measured at a shifted position of ∼922 cm-1. As stated, the sharp amide bands in the ∼1645-1655-cm-1 region arise due to the vibrations of the CO-NH2 moiety, and are highly characteristic of R-helical conformation.43 The occur(42) Miura, T.; Takeuchi, H.; Harada, I. Biochemistry 1988, 27, 88-94. (43) Bandekar, J. Biochim. Biophys. Acta 1992, 1120, 123-143. (44) Overman, S. A.; Thomas, G. J. Biochemistry 1999, 38, 4018-4027.

rence of the amide I band in this study at ∼1658 cm-1 indicated mostly disordered structures. As previously discussed, the increase in intensity indicates either an increase in overall concentration of the scatterer or large-scale reorientation of carbonyl bonds resulting in better overlap of exciting and scattered electromagnetic fields. A decrease in the Phe ∼1003-cm-1 and the Tyr ∼855-cm-1 bands represents a decrease in cellular protein content following apoptotic cell death.37 In that context, the increase of the Phe ∼1003-cm-1 band during quiescent culture may be taken as a spectroscopic indicator that the cardiomyocytes were not undergoing apoptosis. A number of bands in the ∼1100-1400-cm-1 region also appeared to undergo changes during quiescent culture, indicating conformational changes within the sarcoplasm during this study. These changes include the change in intensity of the ∼1123 cm-1 assigned as a pyridine half-ring asymmetry,32 as well as a decrease in the intensity of the ∼1154-cm-1 band indicating the methylene stretch of isoleucine (Ile).44 As previously stated, the band observed at 1448 cm-1 representing a methylene bending mode was used as an internal intensity standard. CONCLUSION Raman spectra were acquired from freshly isolated adult ventricular myocytes without impairing cell function. Conventional confocal Raman spectroscopy may thus be used to monitor changes that result from fast cellular processes such as cardiomyocyte shortening, as well as long-term processes such as the changes associated with quiescent cell culture. A dynamic, timedependent molecular fingerprint of spectral information on these intracellular changes has been generated. The position of the amide I band ∼1658 cm-1, along with its increase over the course of quiescent culture, indicated that disordered domains are the main secondary conformation present and that their concentration increased with time. Changes in intensity of this band during electrically stimulated shortening are believed to result from changes in the orientation of protein scattering centers: the amide I band appeared more intense when their average electric field was aligned with that of the incident light and less intense when not. However, the gradual increase in intensity of this band over the course of quiescent culture indicated a possible increase in disordered structures in the cardiomyocyte sarcoplasm. The fluorescence data showing a decrease in t-tubule density also showed a loss in cellular structure order over a similar time scale. ACKNOWLEDGMENT This work was funded by a grant from the IRC in Bionanotechnology, supported by the EPSRC, MRC, and BBSRC. We acknowledge the help of Lutz Hecht, University of Glasgow in useful discussions.

Received for review November 27, 2006. Accepted March 19, 2007. AC0622476 Analytical Chemistry, Vol. 79, No. 12, June 15, 2007

4587