Synchrotron radiation infrared microspectroscopy of arsenic-induced changes to intracellular biomolecules in live leukemia cells

Vibrational Spectroscopy 53 (2010) 39–44

Contents lists available at ScienceDirect

Vibrational Spectroscopy journal homepage: www.elsevier.com/locate/vibspec

Synchrotron radiation infrared microspectroscopy of arsenic-induced changes to intracellular biomolecules in live leukemia cells Kristie L. Munro a , Keith R. Bambery b , Elizabeth A. Carter c , Ljiljana Puskar d , Mark J. Tobin d , Bayden R. Wood b , Carolyn T. Dillon a,∗ a

Centre for Medicinal Chemistry, School of Chemistry, Building 18, University of Wollongong, Wollongong, NSW 2522, Australia Centre for Biospectroscopy, School of Chemistry, Monash University, Vic 3800, Australia c Vibrational Spectroscopy Facility, School of Chemistry, The University of Sydney, NSW 2006, Australia d Australian Synchrotron, 800 Blackburn Road, Clayton, Vic 3168, Australia b

a r t i c l e

i n f o

Article history: Received 3 September 2009 Received in revised form 10 February 2010 Accepted 11 February 2010 Available online 19 February 2010 Keywords: Arsenic Leukemia Synchrotron FTIR Microspectroscopy Live cells

a b s t r a c t Arsenic trioxide, marketed as TrisenoxTM , successfully cures 60–85% of relapsed acute promyelocytic leukemia sufferers. However, the mechanisms of action remain unclear. In this work, SR-FTIR microspectroscopy of live HL60 cells was used to monitor biomolecular changes that occur during exposure to arsenite (100 ␮M) over a period of 2 h. Importantly, the design of the sample holder enabled the collection of high quality spectra that were not dominated by the ı(OH) mode of water. Significant spectral differences in the live treated cells were observed within 40 min after exposure to the drug. In particular, there was a decrease in the (C O) band at 1742 cm−1 in the spectra of arsenite-treated cells. This is consistent with initial damage to the membrane that leads to later loss of membrane integrity. In addition, there was an initial (40–60 min) decrease in the intensities of the s (PO2 − ) and as (PO2 − ) bands which was attributed to IR “opaqueness” associated with chromatin condensation. In later stage apoptosis (100–120 min) there was an increase in the intensity of these bands which was consistent with DNA fragmentation. The most significant differences (as determined by the second derivative spectra and PCA plots) were observed in the amide I band where the band was centred at 1639 cm−1 (␤-sheet) in the control cells and 1650 cm−1 (␣-helix) in the arsenite-treated cells. The results of this study indicate that the mechanisms of action of arsenite-induced toxicity include alterations to the protein structure and interactions with DNA. Further studies will be performed to verify whether these effects also occur at therapeutic arsenite concentrations (5–10 ␮M). © 2010 Elsevier B.V. All rights reserved.

1. Introduction Arsenic possesses a complex toxicology with severe, acute, chronic and carcinogenic behaviour [1]. Ironically, these toxic properties are pivotal for the success of As therapeutics. For instance, As2 O3 , marketed as TrisenoxTM , is an anti-leukemia agent for the treatment of relapsed acute promyelocytic leukemia (APL) [2]. The successful cure rate of 60–85% of patients who had previously relapsed after conventional therapy, has generated interest in As2 O3 as a treatment for other forms of cancer [2]. New As compounds are also being explored as future therapeutics. To date As

Abbreviations: APL, acute promyelocytic leukemia; FBS, foetal bovine serum; IMDM, Iscove’s modified Dulbecco’s media; PBS, phosphate buffered saline; PCA, principal component analysis; P/S, penicillin/streptomycin; SR-FTIR, synchrotron radiation-Fourier transform infrared. ∗ Corresponding author. Tel.: +61 2 42214930; fax: +61 2 42214287. E-mail address: [email protected] (C.T. Dillon). 0924-2031/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.vibspec.2010.02.004

has been described as a missile with multiple warheads since it has been shown that As can interact with numerous biomolecules including: DNA, tubulin and mitochondrial proteins [3]; however, determining the critical targets for As-induced toxicity towards cancer cells is a necessary precursor to the successful development of this class of drugs. Fourier transform infrared (FTIR) microspectroscopy is an excellent technique for studying changes in the biomolecules associated with biological tissue and cells. Typical spectral features include: amide I ((C O) ∼1650 cm−1 ) and amide II ((CN)/␦(NH) ∼1549 cm−1 ) bands representative of proteins; symmetric (∼1080 cm−1 ) and anti-symmetric PO2 − (∼1232 cm−1 ) bands associated with DNA; C–H stretches of as (CH3 ) (∼2955 cm−1 ) and as (CH2 ) groups (∼2850 cm−1 ), the ester band (C O)ester (1740 cm−1 ) and the C–O–P (∼1080 cm−1 ) band associated with lipids/phospholipids; and C–O stretches (1050 cm−1 ) of carbohydrates [4–7]. One of the most important features of IR spectroscopy is the ability to detect changes in the average protein secondary structure by analysing the components of the

40

K.L. Munro et al. / Vibrational Spectroscopy 53 (2010) 39–44

amide I band; specific secondary structural assignments include: ␣-helices (1662–1645 cm−1 ), ␤-sheets (1637–1613 cm−1 ), turns (1682–1662 cm−1 ) and random coils (1645–1637 cm−1 ) [4,5,8–10]. Consequently, the technique is gaining prominence as a tool capable of diagnosing diseased tissue and monitoring apoptosis [5,7,10,11]. Of particular interest to the area of drug development and design is the implementation of IR spectroscopy to monitor and assess the biomolecular responses to chemotherapeutic agents [12]. Sule-Suso et al. [12] recently reported the use of SR-FTIR microspectroscopy to assess the response of human lung cancer cells to the anti-cancer agent, gemcitabine. Significant changes were observed in the IR spectral region between 1150 cm−1 and 950 cm−1 following treatment with gemcitabine. In particular, there was an increase in the peak at 1080 cm−1 and an increase in the ratio of the peaks at 1080:1050 cm−1 which were attributed to the symmetric vibrations of PO2 groups and the C–O stretches of carbohydrates, respectively [12]. IR spectroscopy has also been used to successfully monitor apoptosis in leukemia cells whereby increases in the lipid content and decreases in cellular DNA were detected. In addition, contributions to the amide I bands showed secondary structure shifts from a predominantly ␤-sheet to unordered protein [13]. Synchrotron radiation (SR) used in combination with FTIR microspectroscopy provides an intense light source (up to 1000× brighter than a globar source) that improves the signal-to-noise at high spatial resolution. This is particularly advantageous in recording IR spectra of single cells where high spatial resolution and, therefore, smaller apertures are essential. Since the size of the focused synchrotron beam is of the order of the cell size it significantly improves the quality of the spectra that can be acquired and reduces the collection time [7,14]. Although an experimental challenge, live cell analysis under aqueous media is desirable as it removes the potential for spectral artefacts resulting from fixation procedures and provides the added advantage of allowing real time analyses. As such SR-FTIR microspectroscopy was used for the first time to analyse live leukemia cells following their exposure to arsenite, the physiological (bioavailable) form of TrisenoxTM . The purpose of the study was to distinguish and identify differences in the intracellular biomolecules (DNA, proteins and membrane lipids) associated with arsenite treatment of leukemia cells in order to identify the targets of As-induced toxicity.

2. Experimental Human APL cells (HL60) were obtained from the ATCC and were grown in 75 cm2 cell culture flasks (Interpath Services) containing growth medium which consisted of Iscove’s modified Dulbecco’s media (IMDM, Invitrogen) supplemented with foetal bovine serum (FBS, 20%, In Vitro Corporation) and penicillin/streptomycin (P/S, 100 U/mL, 100 ␮g/mL, respectively, In Vitro Corporation) in a 37 ◦ C, 5% CO2 , humidified incubator. After reaching confluency, the cells were transferred to 15 mL centrifuge tubes (Interpath Services) and centrifuged at 1000 r.p.m. for 5 min. The cells were then washed in sterile phosphate buffered saline (PBS) and IMDM+P/S (2 mL, without FBS) was added to the centrifuge tubes. Freshly prepared arsenite solution (sodium arsenite, Sigma–Aldrich, 99%) was then added to produce a final As concentration of 100 ␮M. This concentration was chosen since it caused notable toxicity (IC20 ) in the MTT assessment of arsenite following a 4-h exposure period (Supporting Data, Fig. S1). IR spectra of live cells were collected using a demountable liquid cell comprising a 1-mm thick CaF2 window onto which a spacer had been patterned by polymer spin coating and UV photolithography, and an upper 0.5-mm thick CaF2 window [15]. The two windows were held tightly together in a Thermo micro compression cell. To

load the sample into the liquid cell, the resultant cell suspension (1 ␮L) was immediately deposited onto the lower spacer-bearing calcium fluoride disk, and the upper window was placed on top of the sample before closing the compression cell [15]. Excess liquid was expressed through a gap in the microfabricated spacer [15]. Spectra were obtained for ∼20 individual cells after treatment periods of 40 min, 60 min, 100 min and 120 min with arsenite (100 ␮M). Separate calcium fluoride disks were used for HL60 control cells and arsenite-treated cells. The samples were analysed using a Bruker Vertex V80 v Fourier transform infrared (FTIR) spectrometer coupled with a Hyperion 2000 microscope equipped with a liquid nitrogen cooled narrow-band MCT detector at the Australian Synchrotron infrared beamline. Samples were analysed in transmission mode using a 7 ␮m × 7 ␮m aperture with 128 co-added sample scans per cell and 256 co-added background scans (after every 6 cells) using OPUS 6.5 software. Background spectra were obtained from an area (containing IMDM) inside the sample holder which was free from cells and cellular material. Principal component analysis (PCA) was performed using Unscrambler (Version 9.8, Camo, Norway), on all raw spectra that were firstly derivatized (second) (Savitzky–Golay algorithm with 13 smoothing points) and then unit vector normalised. The PCA residual X-variance versus leverage plot for all the spectra was obtained (Supporting Data, Fig. S2) and revealed some outliers which were found to have significant dispersion and Mie scattering contributions. These were removed from the resultant data sets. The resultant spectra for each treatment were averaged using OPUS 6.5 and were presented as raw (no baseline correction) spectra. The spectra were plotted between 3000 cm−1 and 1000 cm−1 due to the presence of water in the aqueous media resulting in a large OH band above 3000 cm−1 which overshadowed many of the other important and relevant peaks. Below 1000 cm−1 the spectral features were difficult to interpret due to the noise. PCA was also performed on spectral regions to identify any significant variation between the data sets. 2.1. Trypan blue exclusion assay HL60 cells were seeded (5 × 105 cells) into 60 mm Petri dishes (Bacto Laboratories) containing growth medium (5 mL) and incubated for 24 h at 37 ◦ C, 5% CO2 in a humidified incubator (Revco Ultima, USA). Following the incubation period the cells were centrifuged at 1000 r.p.m. for 5 min, the supernatant was removed, and the cells were resuspended in IMDM+P/S (control cells) or IMDM+P/S containing 100 ␮M sodium arsenite (treated cells). Following the exposure period (0 min, 40 min, 60 min, 100 min, or 120 min) the cells were washed twice in sterile PBS (1 mL) and were then resuspended in fresh growth media (1 mL) to maintain cell integrity. The cell solution (50 ␮L) was added to trypan blue solution (50 ␮L, 0.5% in PBS, Thermo Electron Corporation) and the cells were scored after 5 min using a haemocytometer. Each assay was performed in triplicate. The number of normal cells (those with intact membranes that contained no stain) and membrane compromised cells (those that did not exclude trypan blue) was plotted as the mean and standard deviation for each time point. Statistical analysis of the data was performed using the ANOVA Tukey–Kramer method whereby a value of p < 0.05 (p = probability) was considered significant [16]. 3. Results Fig. 1 shows the averaged raw spectra obtained from live HL60 cells that were untreated (40 min) or treated with arsenite (100 ␮M,

K.L. Munro et al. / Vibrational Spectroscopy 53 (2010) 39–44

Fig. 1. SR-FTIR averaged raw spectra from live HL60 control cells 40 min after introduction into the sample holder, n = 19 (–); and HL60 cells treated with sodium ); 60 min, n = 20 ( ); arsenite (100 ␮M) for 40 min, n = 20 ( ); and 120 min, n = 15 ( ). The inset shows the 100 min, n = 18 ( normalisation of the spectra over the region (1750–1450 cm−1 ). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

40–120 min). There is evidence of an undulating baseline in all the spectra obtained from the HL60 live cells which is attributed to Mie scattering and a dispersion artefact [17,18]. However, common biomolecular markers, such as the amide I (∼1650–1639 cm−1 ) and amide II (∼1545–1539 cm−1 ) bands (representative of proteins), the anti-symmetric (∼1225 cm−1 ) and symmetric (∼1084 cm−1 ) PO2 − bands (associated with DNA) and the carbonyl ester band (1742 cm−1 , associated with lipids) are clearly discernable in the spectra [4–7]. The most notable feature of the spectra is the similar intensities of the amide I and amide II bands (Fig. 1, inset) which contrasts that seen in fixed cells whereby the amide I band is commonly more intense (approximately 1.5–2×) than the amide II band [7,14,19]. Close examination of the spectrum from the control samples versus the spectra obtained from the arsenite exposed cells over the time course shows differences in the relative intensities of a number of peaks. For instance, the intensity of the as (PO2 − ) and s (PO2 − ) bands at 1225 cm−1 and 1084 cm−1 , respectively, decreased in HL60 cells treated with 100 ␮M sodium arsenite solution (compared to HL60 control cells). In arsenite-treated HL60 cells the intensity of the as (PO2 − ) and s (PO2 − ) increases in cells treated for 100–120 min (compared to cells treated for 40–60 min). Fig. 1(inset) reveals a decrease in the intensity of the (C O)ester (1742 cm−1 ) band (often associated with phospholipids) in HL60 cells treated with sodium arsenite (40–120 min). This suggests that there is a decrease in the amount of observable lipids in arsenitetreated HL60 cells. Fig. 1(inset) also shows a decrease in the intensity of the amide II band with respect to the amide I band following treatment with sodium arsenite (100 ␮M). In addition, the position of the amide II band shifts from 1539 cm−1 in HL60 control cells to 1545 cm−1 in arsenite-treated HL60 cells (Fig. 1) over the 2-h period. The second derivative spectra (Fig. 2) confirm that the amide I band has shifted in the spectra of the HL60 cells treated with 100 ␮M sodium arsenite (40–120 min) compared to control HL60 cells. The amide I band in HL60 control cells is centred around 1639 cm−1 , whereas the amide I band for HL60 cells treated with sodium arsenite (40–120 min) is centred around 1650 cm−1 . The amide I band for HL60 control cells is far broader than what is observed for arsenite-treated HL60 cells indicating that it is potentially comprised of more than one band. This would indicate that the predominant protein structure in HL60 control cells is ␤-sheet, whereas the predominant protein structure in HL60 cells treated

41

with sodium arsenite is ␣-helix and/or random coil. This change is evident in spectra obtained from HL60 cells after the 40 min treatment period indicating that sodium arsenite has a rapid affect on the secondary structure of proteins within HL60 cells. Similarly, the position of the amide II band shifts from 1545 cm−1 to 1547 cm−1 and there is a loss of the shoulder at the lower wavenumber (∼1517 cm−1 ) in arsenite-treated cells. Significant changes are observed in the lineshape, intensity and position of the as (PO2 − ) band. A significant reduction in intensity is observed in the second derivative spectra of the arsenite-treated cells compared to the control cells and there is a shift in the band to ∼1220 cm−1 from 1215 cm−1 . The shoulder at 1237 cm−1 in the control cells is more resolved and distinct in the second derivative spectra of the treated cells. PCA (Fig. 3(a)–(c)) revealed that significant variance exists between the spectra obtained from HL60 control cells and those treated with sodium arsenite (100 ␮M, 40–120 min) for a number of important biomolecules including proteins, lipids, and nucleic acids. The variance was attributed spectral regions containing the amide I and II bands, as well as the (C O)ester , s (PO2 − ) and as (PO2 − ) bands. PCA of the (C O)ester (1758–1671 cm−1 ) region (Fig. 3(a)) shows strong separation along PC1 for HL60 control cells versus arsenite-treated cells. The spectra obtained from arsenitetreated HL60 cells are associated with positive scores in the scores plot. This indicates that they are associated with negative loadings as the PCA was performed on second derivative spectra. The negative loadings correspond to the ester carbonyl band (1742 cm−1 ) and the band associated with DNA base pairing (1715 cm−1 ). Fig. 3(b) shows the results obtained from the PCA performed on the phosphodiester region of the spectrum (1293–1000 cm−1 ). The results show a strong separation along PC1 but unlike Fig. 3(a) the spectra obtained from arsenite-treated cells are associated with negative scores and hence positive loadings. The positive loadings for Fig. 3(b) are associated with the as (PO2 − ) and s (PO2 − ) bands. The same trends are seen in Fig. 3(c) whereby the spectra obtained from arsenite-treated HL60 cells also associated with positive scores and hence negative loadings. The negative loadings are associated with the amide I, amide II, as (PO2 − ) and s (PO2 − ) bands and the positive loadings are associated with the ester carbonyl band. This would indicate that there is an opposite correlation between the lipid bands and the protein/phosphodiester bands in arsenite-treated HL60 cells. PCA revealed no significant difference between the spectra obtained from HL60 cells treated with arsenite following different treatment periods.

Fig. 2. Second derivative SR-FTIR spectra (1800–1000 cm−1 , 13 smoothing points) obtained from HL60 live control cells 40 min after introduction into the sample holder, n = 19 (–) and HL60 live cells treated with sodium arsenite (100 ␮M) for ); 60 min, n = 20 ( ); 100 min, n = 18 ( 40 min, n = 20 ( ); and 120 min, n = 15 ( ). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

42

K.L. Munro et al. / Vibrational Spectroscopy 53 (2010) 39–44

Fig. 3. PCA scores plot and PC1 X-loadings for (a) (C O)ester (1758–1671 cm−1 ), (b) anti-symmetric and symmetric phosphate stretches (1293–1000 cm−1 ) and (c) amide region (1671–1471 cm−1 ) of HL60 control cells (40 min) (A) and HL60 cells treated with sodium arsenite (100 ␮M, 40–120 min) (B). PCA was performed on second derivative (Savitzky Golay, 13 smoothing points), unit vector normalised spectra.

3.1. Trypan blue exclusion assay The results obtained using the trypan blue exclusion assay to assess the condition of the arsenite-treated HL60 cells for the duration (40–120 min) of the IR assay are shown in Fig. 4. The population of cells drops off substantially from that of the control following the 40-min arsenite exposure. However, due to the variability in the cell populations, significant differences in cell populations are not observed until 100- and 120-min exposure periods. While there is more than a doubling in the number of cells that fail to exclude trypan blue for the treated cells (40–120 min) compared with the control cells, this value is not significant (p > 0.05) due to the variation in the cell populations. 4. Discussion Live leukemia (HL60) cells were studied in an effort to analyse cells in an environment which was identical to the conditions in which they were grown. The main advantage of this is that it avoids the incorporation of spectral artefacts that may be associated

Fig. 4. Trypan blue exclusion assay of HL60 cells treated with sodium arsenite (100 ␮M) for 0 min, 40 min, 60 min, 100 min and 120 min. The columns represent the mean of triplicate measurements and the error bars represent the standard deviation from the mean. Cells with intact membranes are represented by white columns (red error bars) and those with compromised membranes are represented by blue columns (blue error bars). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

K.L. Munro et al. / Vibrational Spectroscopy 53 (2010) 39–44

with fixation or other sample preparation procedures. In addition, live cell spectroscopic analysis enables a real time study of the biomolecular changes associated with As exposure. Importantly, however, cell hydration and the surrounding solution can markedly affect the resultant spectra [6]. For instance, the ı(OH) mode of water at 1645 cm−1 is found within the same spectral region as the amide bands and contributes to the intensities of these bands [6]. In addition, the amide I carbonyl group possesses two lone pairs of electrons, and is a likely candidate for H-bonding. This can cause a shift in the position of the amide band. Consequently, a sample holder that enables collection of spectra that are a true representation of the cell but at the same time prevents spectral domination from the ı(OH) water band, is important. The quality of the spectra in Fig. 1 shows that this has been achieved to some extent using the sample holder described [15]. An important difference in these spectra versus those obtained from fixed HL60 cells is the similar intensities of the amide I and amide II bands in the live cells; the amide I band is normally approximately double the amide II band in fixed cells. Moss et al. [20] also reported near-equal intensities of the amide I and II bands in live cell spectra. This has been attributed to the fact that the cell free region used contains more water than the region within a cell, leading to overcompensation of the water band at 1645 cm−1 [20,21]. While mechanisms of correcting for this have been reported (i.e. the qualitative addition of a pure water spectrum to achieve a flat baseline in the spectral region between 3100 cm−1 and 3000 cm−1 ), it is apparent that this is unreliable in microscopic samples [21]. It was concluded by Moss et al. [20] that as long as the same method was used throughout the analysis, this would not void the data comparison. The undulating baseline observed for the HL60 control spectra is characteristic of Mie scattering which results in the observed sinusoidal oscillation in the baseline of the spectrum and is due to dielectric spheres which are known to scatter electromagnetic radiation if the wavelength of the light is comparable to the size of the sphere [18,22]. In this instance the cells are approximately 10 ␮m diameter and contain within them an even smaller nucleus (6–8 ␮m) which makes them an ideal (and unavoidable) candidate for this phenomenon. An additional artefact that can contribute to the baseline is the ‘dispersion artefact’ described by Gardner and coworkers [18]. This is believed to result from scattering by the edges of the cells; although in order to reduce this effect, the aperture size was confined to 7 ␮m in these studies. Nonetheless, careful inspection of the raw spectra (Fig. 1) and second derivative spectra (Fig. 2) reveals important spectral features that can be used to identify the biomolecules in the cells and discern differences between the sample types. The spectral differences observed for the control and arsenite-treated cells are consistent with the fact that there is also a significant decrease (Fig. 4) in the number of viable cells over the exposure period. Firstly, the spectral band observed at 1715 cm−1 , is characteristic of leukemia cells and is believed to be a base-pairing marker of DNA [5]. The decrease in the intensity of this band in the spectra of arsenite-treated HL60 cells indicates a decrease in the amount of detectable DNA. This is consistent with DNA cleavage following the initiation of apoptosis. There is some debate regarding the effects of apoptosis on the intensity of the bands associated with nucleic acids. Studies by Zhou et al. [11] have shown that the intensity of the IR bands assigned to nucleic acids increases during apoptosis while studies by Liu et al. [13] have shown that the intensity of the IR bands assigned to nucleic acids decreases. The change in the intensity of the bands associated with nucleic acids observed in this study may be representative of the progression of apoptosis whereby cells exhibiting nuclear condensation show bands of lower intensity due to low transmission in the infrared of highly condensed areas [22,23]. Non-linear infrared absorption

43

behaviour of the nucleus in a pyknotic (contracted) state associated with increased condensation of DNA has previously been reported [22]. The very high local concentration of DNA is believed to impart an opaqueness of the chromatin, and consequently the absence of DNA in very small nuclei [22]. The small size and the density of the chromatin particles have been implicated in explaining this phenomenon [22]. Cells in later stages of apoptosis after fragmentation of the DNA have been shown to exhibit phosphate bands of higher intensity [23]. These reports [22,23] are consistent with the results of this study, whereby the intensity of the phosphate bands decreased after the 40–60 min treatment period and then subsequently increased after a 100–120 min treatment period. As is expected following arsenite exposure, the spectral results are not consistent with necrosis (versus apoptosis) which is associated with a unique absorption band at 1621 cm−1 [10] and an increase in the intensity of the band at 1740 cm−1 [23]. The decrease in the intensity of the (C O)ester (1741 cm−1 ) band which is often attributed to phospholipids (Fig. 1(inset)) may be indicative of initiation in loss of membrane integrity which is apparent by the increased number of cells that took up trypan blue (Fig. 4) following treatment with sodium arsenite (40–120 min). Finally, the most convincing spectral differences were observed in the contributions to the amide I band providing clear evidence of protein secondary structural changes. These results are also consistent with FTIR microspectroscopy studies of the time dependent exposure of HepG2 cells to arsenite and suggest the probability of direct interaction of As with intracellular proteins [24]. The result is also consistent with our earlier work [25] whereby XAS studies indicated that intracellular As was predominantly Tris–sulphur bound and that this was likely to be protein-bound following treatment with 100 ␮M arsenite (not oxygen- or nitrogen-bound as would be expected following predominance of As–DNA interactions). The results in Figs. 1 and 2 suggest that arsenite treatment results in the loss of the predominance of the ␤-sheet structure [5,8]. The dominance of the contribution at 1650 cm−1 could be attributed to an increase in the relative amount of ␣-helical protein conformation [5,8]. Liu and Mantsch [26] reported that apoptotic leukemia cells also exhibit a change in the protein secondary structure from predominantly ␤-sheet in CEM (T-lymphoblastic cells) control cells to more random coils proteins structures in CEM cells treated with etoposide [13,26]. Studies by Jamin et al. [27] found that in both early and late stage apoptosis the percentage of helical protein increases in relation to the percentage of ␤-sheet protein.

5. Conclusions In this work, the restriction of aqueous solution in the sample holder facilitated a reduction of the dominance of the OH bands in the resultant SR-FTIR spectra. While baseline phenomena, were observed in the control spectra reasonable quality spectra were obtained that showed distinct differences between the control and time dependent arsenite-treated cells. Of major note were the differences in the position of the characteristic amide I band that shifted from 1639 cm−1 , indicative of proteins possessing a predominate ␤-sheet structure (control cells), to 1650 cm−1 (in arsenite-treated cells) which reflects a decrease in the relative ␤sheet structure and dominance of the ␣-helical structure. These results show a clear impact on the protein structure following arsenite treatment, and are consistent with a number of studies that show that As interacts readily with proteins [3,25,28,29]. In addition, there was evidence of changes to DNA (condensation and fragmentation) which is consistent with reports that show that arsenite causes DNA strand breaks at higher concentrations [30]. The successful collection of spectra from live cells is highly valuable for monitoring the effectiveness of anti-leukemia agents such

44

K.L. Munro et al. / Vibrational Spectroscopy 53 (2010) 39–44

as TrisenoxTM , and the ability to distinguish differences between control cells and arsenite-treated cells highlights the validity of performing further studies at therapeutic doses of arsenite (5–10 ␮M). Acknowledgements This research was undertaken on the infrared beamline at the Australian Synchrotron, Victoria, Australia. K.L.M. acknowledges Ian and Jean Simpson and Australian Rotary Health for her Ph.D. scholarship. C.T.D. thanks the University of Wollongong Small Grant Scheme for financial support. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.vibspec.2010.02.004. References [1] WHO, Arsenic and Arsenic Compounds, World Health Organisation, Geneva, 2001. [2] M.T. Rojewski, S. Korper, H. Schrezenmeier, Leuk. Lymphoma 45 (2004) 2387–2401. [3] X. Cai, Y.L. Shen, Q. Zhu, P.M. Jia, Y. Yu, L. Zhou, Y. Huang, J.W. Zhang, S.M. Xiong, S.J. Chen, Z.Y. Wang, Z. Chen, G.Q. Chen, Leukemia 14 (2000) 262–270. [4] A. Barth, Biochem. Biophys. Acta 1767 (2007) 1073–1101. [5] D. Naumann, Appl. Spec. Rev. 36 (2001) 239–298. [6] A. Pevsner, M. Diem, Biopolymers 72 (2003) 282–289. [7] M.J. Tobin, M.A. Chesters, J.M. Chalmers, F.J.M. Rutten, S.E. Fisher, I.M. Symonds, A. Hitchcock, R. Allibone, S. Dias-Gunasekara, Faraday Discuss. 126 (2004) 27–39.

[8] D.M. Blyer, H. Susi, Biopolymers 25 (1986) 469–487. [9] Y. Wang, R.I. Boysen, B.R. Wood, M. Kansiz, D. McNaughton, M.T.W. Hearn, Biopolymers 89 (2008) 895–905. [10] F. Gasparri, M. Muzio, Biochem. J. 369 (2003) 239–248. [11] J. Zhou, Z. Wang, S. Sun, M. Liu, H. Zhang, Biotechnol. Appl. Biochem. 33 (2001) 127–132. [12] J. Sule-Suso, D. Skingsley, G.D. Sockalingum, A. Kohler, G. Kegelaer, M. Manfait, A.J. El Haj, Vib. Spectrosc. 38 (2005) 179–184. [13] K.-Z. Liu, L. Jia, S.M. Kelsey, A.C. Newland, H.H. Mantsch, Apoptosis 6 (2001) 269–278. [14] P. Dumas, G.D. Sockalingum, J. Sule-Suso, Trends Biotechnol. 25 (2006), doi:10.1016/j.tibtech.206.11.002. [15] M.J. Tobin, L. Puskar, R. Barber, E. Harvey, P. Heraud, B.R. Wood, K.R. Bambery, C.T. Dillon, K.L. Munro, Vib. Spectrosc., this issue. [16] GraphPad Software Inc., GraphPad Instat, 1998. [17] J.K. Pijanka, A. Kohler, Y. Yang, P. Dumas, S. Chio-Srichan, M. Manfait, G.D. Sockalingum, J. Sule-Suso, Analyst 134 (2009) 1176–1181. [18] P. Bassan, H.J. Byrne, F. Bonnier, J. Lee, P. Dumas, P. Gardner, Analyst 134 (2009) 1586–1593. [19] M. Diem, M. Romeo, C. Matthaus, M. Miljkovic, L. Miller, P. Lasch, Infrared Phys. Technol. 45 (2004) 331–338. [20] D.A. Moss, M. Keese, R. Pepperkok, Vib. Spectrosc. 38 (2005) 185–191. [21] H. Fabian, P. Lasch, D. Naumann, J. Biomed. Optics 10 (2005) 031103. [22] B. Mohlenhoff, M. Romeo, M. Diem, B.R. Wood, Biophys. J. 88 (2005) 3635–3640. [23] P. Dumas, L. Miller, Vib. Spectrosc. 32 (2003) 3–21. [24] K. Munro, K.R. Bambery, E.A. Carter, L. Puskar, M.J. Tobin, B. Wood, C.T. Dillon, in preparation. [25] K.L. Munro, A. Mariana, A. Klavins, A.J. Foster, B. Lai, Z. Cai, S. Vogt, H.H. Harris, C.T. Dillon, Chem. Res. Toxicol. 21 (2008) 1760–1769. [26] K.-Z. Liu, H.H. Mantsch, J. Mol. Struct. 565–566 (2001) 299–304. [27] N. Jamin, L. Miller, J. Moncuit, W.-H. Fridman, P. Dumas, J.-L. Teillaud, Biopolymers 72 (2003) 366–373. [28] Y.M. Li, J.D. Broome, Cancer Res. 59 (1999) 776–780. [29] Y.-H. Ling, J.-D. Jiang, J.F. Holland, R. Perez-Soler, Mol. Pharmacol. 62 (2002) 529–538. [30] M.J. Mass, A.H. Tennant, B.C. Roop, W.R. Cullen, M. Styblo, D.J. Thomas, A.D. Kligerman, Chem. Res. Toxicol. 14 (2001) 355–361.