Ultrasonic energy as a new tool for fast isotopic 18O labeling of proteins for mass spectrometry-based techniques: Preliminary results

Talanta 76 (2008) 400–406

Contents lists available at ScienceDirect

Talanta journal homepage: www.elsevier.com/locate/talanta

Ultrasonic energy as a new tool for fast isotopic 18 O labeling of proteins for mass spectrometry-based techniques: Preliminary results b ´ R.J. Carreira a , R. Rial-Otero a,1 , D. Lopez-Ferrer , C. Lodeiro a , J.L. Capelo a,∗ a b

REQUIMTE, Departamento de Qu´ımica, Faculdade de Ciˆencias e Tecnologia, Universidade Nova de Lisboa, Quinta da Torre, 2829-516 Monte de Caparica, Portugal Biological Science Division, Pacific Northwest National Laboratory, Richland, WA 99352, USA

a r t i c l e

i n f o

Article history: Received 22 December 2007 Received in revised form 3 March 2008 Accepted 12 March 2008 Available online 21 March 2008 Keywords: Isotopic labeling 18 O Ultrasonic bath Protein quantitation

a b s t r a c t Preliminary results regarding fast isotopic labeling of proteins with 18 O in conjunction with matrix assisted laser desorption ionization time of flight mass spectrometry technique are presented. Similar 16 O/18 O isotopic labeling ratios were found for the overnight procedure (12 h) and the new fast ultrasonic one (30 min) for the BSA, ovalbumin and ␣-lactalbumin proteins. The procedure, however, failed to promote double 18 O isotopic labeling for the proteins, ovalbumin and ␣-lactalbumin. Two different sonication frequencies, 35 and 130 kHz, were studied at two different sonication times of 15 and 30 min, being best results obtained with the procedure at 130 kHz of sonication frequency and 30 min of sonication time. For comparative purposes the overnight isotopic 18 O labeling procedure was done. In addition, the new fast isotopic labeling procedure was also studied without ultrasonication, in a water bath at 60 ◦ C. © 2008 Published by Elsevier B.V.

1. Introduction Protein quantitation is an essential tool for proteomics and systems biology studies, since it helps to understand the function of biological processes, to quantify protein post-translational modifications or to identify diagnosis or prognosis biomarkers, and to develop new drugs [1–3]. Protein quantitation can be done in a relative or absolute manner through mass spectrometry, MS, techniques. MS-based protein quantitative methods can be traced back to the MS stable isotope labeling absolute measurements [4] but, after different improvements in the area, nowadays MS-based strategies for protein relative/absolute quantitation lie mainly into four different approaches, as follows [5]: (i) in the chemical incorporation or “tagging” approach, proteins are derivatizated with a dedicated reagent in a specific site; (ii) in the biological/metabolic incorporation approach, cells are cultured in media enriched in stable isotope-containing amino acids which are incorporated to the cell peptides during the growing process;

∗ Corresponding author. Tel.: +351 212 949 649; fax: +351 212 948 550. E-mail address: [email protected] (J.L. Capelo). URL: http://www.dq.fct.unl.pt/bioscope (J.L. Capelo). 1 University of Vigo, Faculty of Food Science & Technology, Nutrition & Bromatology Group, Analytical & Food Chemistry Department, E-32400 Orense, Spain. 0039-9140/$ – see front matter © 2008 Published by Elsevier B.V. doi:10.1016/j.talanta.2008.03.013

(iii) in the enzymatic incorporation approach, the protein cleavage is done in 18 O-water, thus 18 O is incorporated into the C-terminus of peptides and (iv) in the internal standard approach, a known quantity of an isotopically labeled synthetic peptide is added to the protein digest as internal standard. The use of 18 O-water for isotopic labeling in protein quantitation has its origin in the work of Sprinson and Rittenberg [6]. Enzymatic labeling with 18 O during proteolysis cleavage can be used for relative or absolute quantitation [5,7]. In this process, depending on the enzyme used, one or two oxygen atoms from the solvent (H2 18 O) are incorporated into the peptide C-terminus. For relative quantitation, one set of samples is cleaved in 18 O-water whilst the other is cleaved in 16 O-water. Then, both samples are mixed in equal proportion to ensure that variations due to further sample handling (e.g. losses of peptides) will be paralleled by a proportionate loss of each labeled type (16 O or 18 O), maintaining in this way the ratio between the two. Relative quantitation is achieved by the measurement of the ratios obtained between the intensities of the labeled to unlabeled peptides [8]. Absolute quantitation can also be done with 18 O enzymatic labeling by using standard curves. These curves are done by plotting the protein concentration vs. the ratio obtained between the intensity of a characteristic peptide of the protein and the intensity of an external synthetic peptide used as internal standard, which is also labeled during the protein enzymatic cleavage or later. The internal standard is added in a fixed known quantity to the standards and to the analytical sample. Thus, the intensity ratio of

R.J. Carreira et al. / Talanta 76 (2008) 400–406

the endogenous to synthetic peptide is measured, and the absolute amount of the endogenous peptide can be calculated [9,10]. Enzymatic labeling with 18 O-water has the following advantages: (i) it is of easy labeling; (ii) all peptides present in the sample are labeled and (iii) it requires only the presence of 18 O-water, avoiding extrareagents or synthetic steps. Some drawbacks, however, inherent to this type of labeling need to be overcome. Thus, systematic studies have shown that different types of proteolytic enzymes incorporate different levels of 18 O from water during digestion [5]. Ideally, the incorporation of two 18 O (18 O2 ), that is, a mass shift of 4 Dalton (Da), would be the minimum needed to obtain appreciable m/z changes in the mass spectrum between samples, avoiding, for instance, the 13 C interference or isotopic peak overlapping. The incorporation of a single 18 O (18 O ), yields a mass increment between each sample of only 1 2 Da, which is not adequate to produce an appreciable change in m/z, especially for multiply charged species. Furthermore, the procedure for isotopic labeling is a time-consuming approach which can take between 12 and 48 h. Ultrasound energy has been used in chemistry with the aim to speed up kinetics of chemical reactions [11,12]. Remarkably, ultrasonic energy has been recently reported as a tool for the acceleration of enzymatic protein cleavage from overnight to few minutes [13,14]. However, the potential of ultrasound energy to speed up the 18 O enzymatic labeling process and to increase the ratio of 18 O incorporation remains unknown. In this work, we report our preliminary results regarding the acceleration of 18 O isotopic labeling joining the enzymatic protein digestion process with ultrasonic energy. With this aim, 18 O labeling with ultrasonic energy provided by an ultrasonic bath at 60 ◦ C is studied. The influence of the ultrasonic frequency is also studied by using 35 and 130 kHz ultrasonic transducers. In addition, for comparative purposes the overnight labeling and labeling in water at 60 ◦ C are also done. 2. Experimental 2.1. Apparatus Protein digestion/labeling was done in safe-lock tubes of 0.5 ml from Eppendorf (Hamburg, Germany). A minicentrifuge, model Spectrafuge-mini, from Labnet (Madrid, Spain), and a minicentrifuge-vortex, model Sky Line, from ELMI (Riga, Latvia) were used throughout the sample treatment, when necessary. Milli-Q natural abundance (H2 16 O) water was obtained from a SimplicityTM 185 from Millipore (Milan, Italy). An ultrasonic bath, model Transsonic TI-H-5, from Elma (Singen, Germany) with control of temperature and amplitude was used. 2.2. Materials and reagents Bovine serum albumin, BSA (66 kDa, >97%), ovalbumin (45 kDa), ␣-lactalbumin (14.4 kDa, ≥85%) and trypsin (proteomics grade) used for all experiments, were purchased from Sigma (Steinheim, Germany). Reduction and alkylation of BSA were carried out, respectively, with dl-dithiothreitol (DTT, 99%) and iodoacetamide (IAA) from Sigma. The following reagents were used during sample digestion/labeling: ammonium bicarbonate buffer (AmBic, pH 8.5, ≥99.5%) and formic acid (FA, ∼98%) from Fluka (Buchs, Switzerland); H2 18 O (95 atom%) from ISOTECTM (Miamisburg, USA). ␣-Cyano-4-hydroxycinnamic acid (␣-CHCA, ≥99.0%), ace-

401

tonitrile (99.9%) and trifluoroacetic acid (TFA, 99%) were from Fluka, ¨ (Seelze, Germany), respectively. Sigma–Aldrich and Riedel-de Haen ProteoMassTM Peptide MALDI-MS calibration kit (MSCAL2) from Sigma was used as mass calibration standard for MALDI-TOFMS. 2.3. Sample treatment A stock solution of BSA, ovalbumin and ␣-lactalbumin (100 pmol/␮l) was prepared in AmBic (100 mM) using natural abundance water. Reduction was performed with DTT (10 mM) at 37 ◦ C for 1 h and alkylation was done with IAA (50 mM) at room temperature (RT) for 45 min in the darkness. Then aliquots (10 ␮l) of the protein solution were taken and diluted to 100 ␮l with AmBic (100 mM) prepared in natural abundance water or in 95% 18 Oenriched water. Trypsin (2%, v/v) was added to these solutions to a final concentration of 0.47 pmol/␮l. The substrate to enzyme ratio was 20:1 (mol/mol). Different digestion and labeling procedures were tested: (i) overnight/labeling (12 h) digestion at 37 ◦ C; (ii) digestion/labeling at 60 ◦ C in an ultrasonic bath (70% sonication amplitude and 35 kHz sonication frequency) for variable times of 15 and 30 min; (iii) digestion/labeling at 60 ◦ C in an ultrasonic bath (70% sonication amplitude and 130 kHz sonication frequency) for variable times of 15 and 30 min; (iv) digestion/labeling at 60◦ C for variable times of 15 and 30 min. To stop enzymatic digestion/labeling reactions, 5 ␮l of formic acid 50% were added. A comprehensive scheme of the sample treatment is presented in Fig. 1. 2.4. MALDI-TOF-MS analysis Prior to MALDI-TOF-MS analysis, the samples were mixed in a 1:1 ratio with the matrix solution of ␣-CHCA (10 ␮g/␮l) prepared in Milli-Q water/acetonitrile/TFA (1 ml + 1 ml + 2 ␮l). Then, 1 ␮l of each sample was hand-spotted onto a stainless steel well plate of a MALDI-TOF-MS and allowed to dry. MALDI mass spectra were obtained with a Voyager DE-PRO Biospectrometry Workstation model from Applied Biosystems (Foster City, USA), equipped with a nitrogen laser radiating at 337 nm. Measurements were done in the reflector positive ion mode, with a 20 kV of accelerating voltage, 75.1% of grid voltage, 0.002% of guide wire and a delay time of 100 ns. Two close external calibrations were performed with the monoisotopic peaks of the Bradykinin, Angiotensin II, P14R and ACTH peptide fragments (m/z [M+H]+ : 757.3997, 1046.5423, 1533.8582 and 2465.1989, respectively). 250 laser shots were summed per spectrum. The spectra analysis was done with the Data ExplorerTM software (version 4.0) from Applied Biosystems. The following search engines were used to identify the obtained peptide fingerprints: MASCOT [http://www.matrixscience.com/ search form select.html] and PROTEIN PROSPECTOR [http:// prospector.ucsf.edu/]. Search parameters: (i) SwissProt. 2006 Database; (ii) molecular weight (MW) of protein: all; (iii) one missed cleavage; (iv) fixed modifications: carbamidomethylation (C); (v) variable modifications: oxidation (M), 18 O1 and 18 O2 label (C-term); (vi) peptide tolerance up to 175 ppm. If the protein identification score is located out of the random region and the protein analyzed scores first, then a match is considered successful.

402

R.J. Carreira et al. / Talanta 76 (2008) 400–406

in order to test the correct applicability of this function, the mathematical algorithm for deconvolution described by Yao et al. was also used in the first steps of this work [16]. 3. Results and discussion 3.1. The isotopic labeling and the deconvolution problem

Fig. 1. Comprehensive scheme for the isotopic labeling sample treatment.

2.5. Isotopic peak deconvolution Isotopic peak deconvolution was done using the deisotope function of the Data ExplorerTM software (version 4.0) from Applied Biosystems. This function is an advanced peak filtering method that uses a deisotoping algorithm to determine the relative abundance of multiple components with overlapping isotope distributions [15]. Thus, the deisotope function allows reducing a spectrum to a centroided plot by deconvoluting the monoisotopic peaks from the peak list. For each peak in a spectrum, the software inspects the peak list for the higher theoretical masses and areas associated with additional expected peaks in a theoretical isotopic cluster, relative to the peak in question. Moreover, for comparative purposes

Fig. 2 presents the MALDI-TOF-MS spectra of the peptide fragment (YLYEIAR)H+ (m/z = 927 Da) obtained from the tryptic digestion of BSA and acquired under the conditions briefly described in the caption of Fig. 2 (for further details refer to Section 2.3). The peptide (YLYEIAR)H+ is used throughout this manuscript for explanation purposes because the best isotopic labeling was obtained with this peptide. As it was explained in Section 1, the incorporation of one 18 O (18 O1 ) yields a mass increment for each isotopic peak of 2 Da whilst the incorporation of two 18 O (18 O2 ) yields a mass increment of 4 Da (see Fig. 2a and b). To avoid isotopic peak overlapping a 100% 18 O2 labeling should be achieved. The difference in the isotopic patterns due to variable labeling and its consequence for the spectra interpretation can be easily explained through Fig. 2, spectra a–c. Spectrum a, Fig. 2, corresponds to the BSA overnight digestion in 16 O-water and, as it can be seen, the isotopic pattern match well with the natural isotopic distribution, as showed in Table 1, with the following main m/z peaks being obtained: 927, 928 and 929 Da. Ideally, as explained before, the isotopic labeling should incorporate two 18 O, giving a mass increment of 4 Da for each isotopic peak. Therefore, the following main m/z peaks should be expected 931 Da (=927 + 4), 932 Da (=928 + 4) and 933 Da (=929 + 4) for a 100% of 18 O2 labeling. Experimentally, however, this does not occur because the yield of double oxygen incorporation is not 100%. This is the reason why spectrum b, Fig. 2, presents m/z peaks at 929 and 930 Da, which should not appear for a 100% of 18 O2 labeling. The peak at 929 Da observed in Fig. 2b is originated by the incorporation of one 18 O to the peptide (YLYEIAR)H+ (927 + 2) plus the 929 Da peak contribution coming from the unlabeled peptide. The peak at 931 Da has contributions coming from one single 18 O incorporation (929 + 2) and from double 18 O incorporation to the peptide (YLYEIAR)H+ (927 + 4). Since there is no way to know the percentage of peptides that were unlabeled, single labeled or double labeled, it is necessary to use complicated mathematical procedures to deconvolute the spectrum [7–9]. Further understanding of this problem is taken from the other spectra present in Fig. 2, where it can be seen different intensities for the peaks of m/z 929 and 931, for the same sample and the same concentration, as a result of the different conditions in which the digestion and the isotopic labeling were done, thus indicating different degrees of labeling (single or double) as a function of the sample treatment. 3.2. Influence of the sample treatment on the 18 O labeling (16 O/18 O) The proteolytic 18 O labeling method can be done in two different ways, as explained below. The first approach, named as

Table 1 Theoretical vs. experimental isotopic distribution for the tryptic peptide fragment (YLYEIAR)H+ from BSA digestion Mass (m/z) [M+H+ ]

Theoretical isotopic distributiona

Experimental isotopic distributionb

Mass (m/z) after 18 O1 incorporation

Mass (m/z) after 18 O2 incorporation

927 928 929 930

100.0 53.8 16.6 3.7

100 54.1 15.6 3.4

929 930 931 932

931 932 933 934

a b

Isotopic distribution calculated with the isotopic calculator function of the Data ExplorerTM software (version 4.0) from Applied Biosystems. Experimental ratios obtained for overnight digested BSA samples. Values were acquired with a MALDI-TOF-MS system in the reflector positive ion mode (n = 2).

R.J. Carreira et al. / Talanta 76 (2008) 400–406

403

Fig. 2. Reflector positive ion mode MALDI-TOF mass spectra of the peptide fragment (YLYEIAR)H+ obtained in the tryptic digest of BSA. (a) Overnight digestion in 100% H2 16 O buffer solution at 37 ◦ C (ca. 12 h); (b) overnight digestion in 95 atom% H2 18 O buffer solution at 37 ◦ C (ca. 12 h); (c) ultrasonic bath digestion in 95 atom% H2 18 O buffer solution at 60 ◦ C (35 kHz frequency; 70% amplitude; 30 min); (d) ultrasonic bath digestion in 95 atom% H2 18 O buffer solution at 60 ◦ C (130 kHz frequency; 70% amplitude; 15 min); (e) ultrasonic bath digestion in 95 atom% H2 18 O buffer solution at 60 ◦ C (130 kHz frequency; 70% amplitude; 30 min) and (f) water bath digestion in 95 atom% H2 18 O buffer solution at 60 ◦ C (30 min).

“direct labeling”, incorporates the 18 O at the same time that the enzymatic cleavage is done [8,9]; this is, the enzymatic cleavage is performed in 18 O-water, normally at the pH recommended by manufacturers (pH 8.5) to obtain the best efficiency of the enzyme. The second approach, named as “decoupling procedure”, was first reported by Yao et al. [17], and briefly, it decouples the labeling process into two steps, as follows: first, the enzymatic digestion is done in 16 O-water, at the recommended pH for the best enzymatic efficiency and then, after 16 O-water evaporation, the labeling process takes place in 18 O-water at a lower pH (5–6). The 18 O labeling decoupling process is based on recent literature, where it is suggested that optimum pH for almost complete two 18 O labeling should be shifted toward acidic pHs, compared to the pH range of the enzymes highest activities [7]. Although, the decoupling procedure has the advantage of separating the digestion and labeling steps, allowing their conditions to be optimized separately, in our experiments we decided to use direct labeling as mentioned above, since it was expected a smaller technical variation. In addition, direct labeling is done in one single step facilitating on-line approaches for mass spectrometry. The 18 O labeling process can be divided into two chemical reactions, as follows: (i) first reaction: amide bond cleavage RC16 ONHR  + H2 18 O → [RC16 O18 O]− + [H3 NR  ]+ (ii) second reaction: carboxyl oxygen exchange [RC16 O18 O]− + H2 18 O → [RC18 O18 O]− + H2 16 O

The first experiments were done to study the influence of ultrasonic energy on the total enzymatic labeling of peptides with 18 O. The ratio 16 O/18 O refers to the amount of 18 O incorporated in the labeling process, no matter the type of labeling (single or double). Table 2 shows the 16 O/18 O ratios of different peptides obtained from the tryptic digestion of BSA after performing the different sample treatments studied in this work.

Concerning the overnight protocol (12 h), the yield of 18 O incorporation all the peptides shown in Table 2 was higher than 95% (16 O/18 O ratios lower than 0.05). Regarding 18 O incorporation with the treatment with ultrasound at 130 kHz sonication frequency at 60 ◦ C and the treatment with the water bath at 60 ◦ C (no ultrasonication was used), data in Table 2 show that similar results, a labeling efficiency of ca. 90% (16 O/18 O ratios lower than 0.13), between both methods were obtained. In addition, the yields of incorporation were close to the ones obtained for the overnight protocol, being the labeling slightly worst for the higher peptides. However the labeling efficiency was still ≥90%. As far as the treatment with 35 kHz sonication time at 60 ◦ C concerns, data showed in Table 2 seems to suggest a dependence of labeling efficiency with the sonication frequency. Thus the 18 O incorporation was worst that the one obtained with the same time and the same temperature, but with a sonication frequency of 130 kHz. On the overall, for the aforementioned treatments, the labeling efficiency was not equally obtained for all peptides. It seems that the labeling yield was dependent of the peptide size. Thus, as showed in Table 2, labeling efficiencies lower than 90% (16 O/18 O ratios higher than 0.11) were obtained for the peptides (RHPEYAVSVLLR)H+ , (LEGYGFQNALIVR)H+ and (KVPQVSTPTLVEVSR)H+ , with peak masses of 1439, 1479 and 1639 m/z, respectively; whilst labeling efficiencies higher than 90% (16 O/18 O ratios lower than 0.11) were observed for the peptides (YLYEIAR)H+ and (ALKAWSVAR)H+ with masses 927 and 1001 m/z, respectively. Interestingly, this problem was not observed when the labeling process was performed overnight. This fact could suggest that the times (15 and 30 min) selected to speed up the enzymatic digestion and labeling with ultrasound or heating at 60 ◦ C were not enough to increase the reaction rates of the enzymatic process. A further explanation can be suggested taking into account that the yields of labeling were similar for the ultrasonic process at 130 kHz sonication frequency at 60 ◦ C and for the water bath digestion at 60 ◦ C. So, although different authors have suggested that ultrasonication can accelerate the enzymatic reactions, little attention has been given to the type of device used to perform such treatment, that can be an ultrasonic bath, such as in this work, or

404

R.J. Carreira et al. / Talanta 76 (2008) 400–406

Table 2 16 O/18 O ratios of different peptides obtained from the tryptic digestion of BSA, ovalbumin and ␣-lactalbumin (900 pmol each protein), in the presence of 95% H2 18 O Peptide fragment

[M+H]+ (m/z)

Overnight

(YLYEIAR)H+

927.49

0.01 ± 0.02

(ALKAWSVAR)H+

1001.59

a

(RHPEYAVSVLLR)H+

1439.82

0.04 ± 0.01

1479.80

0.04 ± 0.01

(KVPQVSTPTLVEVSR)H

1639.94

0.03 ± 0.02

Ovalbumin

(VYLPR)H+ (HIATNAVLFFGR)H+ (GGLEPINFQTAADQAR)H+ (ELINSWVESQTNGIIR)H+ (LYAEERYPILPEYLQCVK)H+

647.39 1345.74 1687.84 1858.97 2284.17

0.02 ± 0.02 0.03 ± 0.01 0.07 ± 0.06 0.19 ± 0.17 0.08 ± 0.03

␣-Lactalbumin

(CEVFR)H+ (VGINYWLAHK)H+ (EQLTKCEVFR)H+ (ILDKVGINYWLAHK)H+

710.33 1200.65 1309.66 1669.94

0.03 ± 0.00 0.04 ± 0.03

Protein

BSA

(LEGYGFQNALIVR)H+ +

a

0.09 ± 0.07

US bath 35 kHz

US bath 130 kHz

Water bath

0.07 ± 0.01 0.05 ± 0.00 0.08 ± 0.02 0.06 ± 0.03 0.11 ± 0.02 0.11 ± 0.02 0.31 ± 0.05 0.23 ± 0.06 0.11 ± 0.03 0.08 ± 0.03

0.03 0.02 0.03 0.06 0.05 0.06 0.08 0.11 0.08 0.08

± ± ± ± ± ± ± ± ± ±

0.01 0.01 0.04 0.05 0.02 0.01 0.01 0.04 0.00 0.02

b

0.09 0.13 0.11 0.15 0.15

± ± ± ± ±

0.00 0.01 0.01 0.01 0.02

b

30

b

0.14 0.14 0.14 0.17

± ± ± ±

0.01 0.03 0.01 0.01

b

30

0.04 0.03 0.06 0.05 0.08 0.07 0.12 0.13 0.11 0.06

± ± ± ± ± ± ± ± ± ±

0.00 0.03 0.01 0.03 0.01 0.01 0.02 0.01 0.02 0.05

Sonication time (min) 15 30 15 30 15 30 15 30 15 30

Different labeling methods were used in this experimental: overnight (12 h); ultrasonic bath (35 kHz; 70% amplitude, 60 ◦ C, 15 and 30 min); ultrasonic bath (130 kHz; 70% amplitude, 60 ◦ C, 15 and 30 min); water bath (60 ◦ C, 15 and 30 min). Values were acquired with a MALDI-TOF-MS system in the reflector positive ion mode (n = 2). a Peptide not present in the spectra. This peptide has a missed cleavage. b Experimental was not done.

an ultrasonic probe [18,19]. Thus, for a given volume of solution, the ultrasonic intensity obtained with an ultrasonic bath is 100 times lower than the ultrasonic intensity obtained with an ultrasonic probe [20]. Therefore, the low yields of 18 O labeling observed for the ultrasonication treatment for the peptides of higher masses could be explained taking into account the low sonication intensity of the ultrasonic bath. Thus, in our case, the 16 O/18 O labeling yields observed for the sample treatment with the ultrasonic bath could be linked to the temperature at which the labeling was done rather than to the ultrasonication effects. In fact, the labeling ratios for the treatment with ultrasonic bath at 60 ◦ C and 130 kHz of sonication frequency and with the heating treatment in water at 60 ◦ C without sonication are similar for all BSA peptides. Nevertheless, as we will see below, the 18 O1 /18 O2 labeling efficiencies obtained with both methods were completely different. Another interesting finding deal with the different yields of isotopic incorporation obtained as a function of the sonication frequency used, as explained above. The ultrasonication bath employed in this work has the possibility of performing two different ultrasonication frequencies of 35 and 130 kHz. As it can be seen in Table 2, the results obtained with the ultrasonication bath at the frequency of 130 kHz and 60 ◦ C are similar to the overnight protocol for all the BSA peptides observed, but the labeling achieved with the 35 kHz sonication frequency was low, specially for the peptide (LEGYGFQNALIVR)H+ . The explanation could be as follows: when ultrasonic energy is provided to a liquid media, an effect known as cavitation occurs. This effect is responsible for the generation of small gas bubbles, which grow in successive cycles, according to the frequency of ultrasound, and when they reach an unstable size they violently collapse. The temperature and pressure near the collapse point can reach values up to 5000 ◦ C and 1000 atm, respectively [21]. This energy is transmitted to the liquid media and the mass transfer processes are enhanced in a more effective manner than when only heat is used. Two types of cavitation can be produced by ultrasound: transient cavitation and stable cavitation. Most of the physical and chemical effects produced by ultrasonication are linked to the collapse of transient cavitation bubbles [21], which are dominant at low frequencies, i.e. 35 kHz. However, as the frequency increases, so it does the fraction of stable cavitation bubbles formed. As the transient cavitation bubbles to stable cavi-

tation bubbles ratio changes, the sonochemical effects are altered. We, therefore, hypothesize that the worst performance of the enzymatic labeling at lower frequencies (i.e. 35 kHz) is directly linked with the increasing number of transient cavitation bubbles formed. 3.3. Influence of the sample treatment on the single to double 18 O labeling The cause of the major variability of the 18 O labeling technique applied to quantitative proteomics is the variable incorporation of 18 O atoms into peptides, and this variability is directly linked to the second 18 O incorporation [7–9]. This is due to the fact that the efficiencies of the two reactions involved in the isotopic labeling with 18 O, the amide bond cleavage and the carboxyl oxygen exchange, are different. The first reaction incorporates one 18 O in one cycle of the enzymatic reaction, whilst the second reaction needs at least five cycles to have a yield of 18 O2 incorporation of ca. 98.5% [7]. This reaction is extremely slow under the conditions commonly used leading to variable exchange within the time frame of the proteolytic reaction [17]. The cavitation phenomena generated by ultrasonic energy enhances both of the labeling reactions, but in the second reaction the forward and backward processes are probably both accelerated. In this way it must be stressed out that dedicated literature suggest that ultrasonication in short times can accelerate enzymatic reactions, whilst in longer times it seems that enzymatic inactivation occurs. Thus, Talukder et al. [18] reported an increment in the maximum reaction rate (Vmax ) and unalteration for the Michaelis constant (Km ) when the hydrolytic activity of lipase was enhanced by ultrasonication. Therefore, it is critical to control the carboxyl oxygen exchange reaction (second reaction) as the key strategy in overcoming variability. Table 3 shows data regarding single to double labeling ratios (18 O1 /18 O2 ) for the BSA protein obtained with the different sample treatments studied. As it can be seen, the (18 O1 /18 O2 ) labeling follow the same pattern that the 16 O/18 O incorporation; this means that the best results were obtained with the overnight protocol (18 O1 /18 O2 ratio lower then 0.35, that means 18 O2 incorporation higher than 75%) and, from all the other sample treatments studied, only the acceleration with the ultrasonic bath at 130 kHz sonication frequency at 60 ◦ C showed to be a method of relative success. It is remarkable the fact that the

R.J. Carreira et al. / Talanta 76 (2008) 400–406

405

Table 3 18 O1 /18 O2 ratios of different peptides obtained from the tryptic digestion of BSA, ovalbumin and ␣-lactalbumin (900 pmol each protein), in the presence of 95% H2 18 O Protein

BSA

Peptide fragment

[M+H]+ (m/z)

Overnight

(YLYEIAR)H+

927.49

0.32 ± 0.04

(ALKAWSVAR)H+

1001.59

a

(RHPEYAVSVLLR)H+

1439.82

0.29 ± 0.05

1479.80

0.34 ± 0.04

(KVPQVSTPTLVEVSR)H

1639.94

0.30 ± 0.00

(VYLPR)H+ (HIATNAVLFFGR)H+ (GGLEPINFQTAADQAR)H+ (ELINSWVESQTNGIIR)H+ (LYAEERYPILPEYLQCVK)H+

647.39 1345.74 1687.84 1858.97 2284.17

0.36 ± 0.06 0.45 ± 0.07 0.52 ± 0.06 1.02 ± 0.93 0.90 ± 0.87

(CEVFR)H+ (VGINYWLAHK)H+ (EQLTKCEVFR)H+ (ILDKVGINYWLAHK)H+

710.33 1200.65 1309.66 1669.94

0.31 ± 0.03 0.39 ± 0.21

(LEGYGFQNALIVR)H+ +

Ovalbumin

␣-Lactalbumin

a

5.98 ± 0.43

US bath 35 kHz

US bath 130 kHz

1.30 ± 0.16 0.80 ± 0.23 1.50 ± 0.33 0.83 ± 0.18 4.33 ± 0.61 2.11 ± 0.84 3.51 ± 0.27 2.59 ± 0.30 24.21 ± 2.17 21.07 ± 17.47

0.41 0.42 0.40 0.32 0.77 0.76 1.15 1.21 3.61 3.06

b

± ± ± ± ± ± ± ± ± ±

0.05 0.01 0.03 0.09 0.07 0.01 0.04 0.26 0.18 0.13

2.12 ± 0.30 3.89 ± 0.37 3.18 ± 0.48

Water bath

Sonication time (min)

0.80 ± 0.08 0.51 ± 0.19 0.12 ± 0.38 0.55 ± 0.07 1.42 ± 0.02 0.84 ± 0.48 1.33 ± 0.23 1.09 ± 0.21 7.77 ± 5.37 2.63 ± 1.03

15 30 15 30 15 30 15 30 15 30

b

30

b

30

c c

b

19.38 ± 3.66 32.25 ± 1.34 37.53 ± 13.50 c

Different labeling methods were used in this experimental: overnight (12 h); ultrasonic bath (35 kHz; 70% amplitude, 60 ◦ C, 15 and 30 min); ultrasonic bath (130 kHz; 70% amplitude, 60 ◦ C, 15 and 30 min); water bath (60 ◦ C, 15 and 30 min). Values were acquired with a MALDI-TOF-MS system in the reflector positive ion mode (n = 2). a Peptide not present in the spectra. This peptide has a missed cleavage. b Experimental was not done. c Double labeled peptide was not identified in the MALDI mass spectra.

heating at 60 ◦ C although could promote a 16 O/18 O labeling similar to the ultrasonic process at 130 kHz sonication frequency at 60 ◦ C, failed, however, in promoting double 18 O incorporation. As far as the size of the peptide concerns, the double labeling with the overnight protocol or with the ultrasonic bath at 60 ◦ C and 130 kHz showed to be the most robust procedures, with similar results: the peptides (YLYEIAR)H+ and (ALKAWSVAR)H+ (m/z = 927 and 1001, respectively) were labeled with a yield higher than 70% (18 O1 /18 O2 ratio lower then 0.43). Nevertheless, for the peptides (RHPEYAVSVLLR)H+ , (LEGYGFQNALIVR)H+ and (KVPQVSTPTLVEVSR)H+ (m/z = 1439, 1479 and 1639, respectively) a double labeling efficiency of about 70% was achieved with the overnight protocol whilst with the ultrasonic protocol (130 kHz and 60 ◦ C) lower double labeling efficiency was obtained and also this efficiency decreased as the peptide size increased. Other authors have also stressed that different peptides may exhibit different relative labeling efficiencies [7]. The reasons for the aforementioned fact, however, are not well understood yet. For instance, Mirgorodskaya et al. [9] reported similar 18 O1 /18 O2 ratios for peptide fragments with masses comprised between 1050.5 and 1385.6 m/z from a RNase tryptic digest, whilst Stewart et al. [8] have found 18 O /18 O ratios comprised between 25% and 80%, as a function of 1 2 the peptide-type. 3.4. Application to further proteins In order to assess the efficiency of the ultrasonic procedure at 60 ◦ C and 130 kHz of sonication frequency to further proteins, the ovalbumin the ␣-lactalbumin proteins were submitted to the above-mentioned procedure for a time of 30 min. In addition the overnight procedure was also done for comparative purposes. Results are presented in Tables 2 and 3 for 16 O to 18 O labeling and single to double 18 O labeling, respectively. Regarding ovalbumin, the efficiency of the 16 O/18 O labeling, was similar with both procedures as it can be seen in Table 2. Thus, for the overnight procedure the labeling efficiency was comprised between 84% and 98% (16 O/18 O ratio comprised between 0.19 and 0.02), whilst for the ultrasonic procedure the incorporation was comprised between 87% and 92% (16 O/18 O ratio comprised between

0.15 and 0.09). That means a reduction in the sample time needed to perform the isotopic labeling of ca. 12 h when ultrasonication was used. However, the ultrasonication treatment failed in promoting the double oxygen incorporation as showed in Table 3, were it can be seen that for the overnight procedure the double incorporation was comprised between 50% and 73%, (18 O1 /18 O2 ratio between 1 and 0.36) whilst for the ultrasonic process the double incorporation was comprised between 20% and 32% (18 O1 /18 O2 ratio between 4 and 2.1). Concerning protein ␣-lactalbumin, for the overnight procedure the 16 O/18 O labeling efficiency, was comprised between 65% and 97% (16 O/18 O ratio comprised between 0.54 and 0.03), whilst for the ultrasonic procedure the incorporation was comprised between 85% and 88% (16 O/18 O ratio comprised between 0.17 and 0.14). However, the ultrasonication treatment was almost non-effective in promoting the double oxygen incorporation. Thus, the best double incorporation was for the peptide fragment (CEVFR)H+ , and it was as poor as 5% (18 O1 /18 O2 ratio higher than 19). The double incorporation for the same peptide with the overnight protocol was 76% (18 O1 /18 O2 ratio = 0.31). 4. Future prospects On the overall, the possibility of speeding up 18 O isotopic labeling from 12–48 h to 30 min using ultrasonication with bath and 130 kHz sonication frequency is a reliable approach that deserves further investigation due to the importance of protein quantitation for biomarker discovery and clinical diagnosis either in a relative or absolute manner. It is clearly stated from the data here presented that ultrasonication can accelerate the 16 O/18 O labeling, but it seems to fail in promoting the double 18 O incorporation (18 O2 ). However this failing is selective and it depends on the protein. In addition, for the same protein, when the ultrasonication is used, the double 18 O incorporation is peptide-selective. Therefore, further research needs to be done focusing on the following variables: regarding ultrasonic energy, it is necessary to establish if the ultrasonic frequency is a variable of main importance linking it with other ultrasonic devices different from the ultrasonic bath, such as the sonoreactor or the ultrasonic probe. Preliminary results

406

R.J. Carreira et al. / Talanta 76 (2008) 400–406

obtained in our laboratory seem to suggest that high differences can be obtained in the 18 O incorporation as a function of the ultrasonic device used to speed up the labeling process. Concerning proteins, the sample treatment must be studied linking the ultrasonic efficiency to the physical and chemical characteristics of the protein. In addition, special attention must be paid to the 18 O incorporation as a function of the type of peptide. 5. Conclusions Our preliminary results demonstrate that ultrasonic energy has a great potential to be used to enhance the enzymatic isotopic labeling of peptides with 18 O. The time needed to perform the isotopic labeling can be reduced from 12–48 h to 30 min with similar results to the ones obtained with the overnight protocol, depending on the protein. Similar 16 O/18 O ratios were found for BSA, ovalbumin and ␣lactalbumin proteins between the ultrasonic procedure and the overnight one, with a total time used to perform the isotopic labeling of 30 min. However, the 18 O1 /18 O2 labeling ratios obtained were considered acceptable only for the BSA protein. The fact that sonication frequency (35 kHz vs. 130 kHz) had been found critical for the good performance in accelerating the labeling process, open new ways in re-thinking ultrasonic applications in the analytical laboratory for proteomic purposes. It must be taken into account that most ultrasonic devices used at present, work at low frequencies such as 20, 22, 30 or 35 kHz. Further research is to date being done in our laboratory to elucidate the key parameters on this new methodology, including the use of different ultrasonic devices with different frequencies of work. Acknowledgments R. Rial-Otero and R.J. Carreira acknowledge the postdoctoral grant SFRH/BPD/23072/2005 and the doctoral grant

SFRH/BD/28563/2006, respectively, of FCT (Science and Technological Foundation) from Portugal. Dr. J.L. Capelo acknowledges the MALDI-TOF-MS service of the Chemistry Department of the New University of Lisbon (http://www.dq.fct.unl.pt/maldi) for their helpful assistance and valuable suggestions. The research findings here reported are protected by international laws under patent pending PCT/IB2006/052314 and PT 103 303. FCT is also acknowledged for financial support under the project POCI/QUI/55519/2004 FCT-FEDER. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21]

M. Hamdam, P.G. Righetti, Mass Spectrom. Rev. 21 (2002) 287–302. S. Sechi, Y. Oda, Curr. Opin. Chem. Biol. 7 (2003) 70–77. W.W. Wu, G. Wang, S.J. Baek, R.F. Shen, J. Proteome Res. 5 (2006) 651–658. A.P. De Leenheer, L.M. Thienpont, Mass Spectrom. Rev. 11 (1992) 249–307. M. Bantscheff, M. Schirle, G. Sweetman, J. Rick, Bernhard Kuster, Anal. Bioanal. Chem. 389 (2007) 1017–1031. D.B. Sprinson, D. Rittenberg, Nature 167 (1951) 484. M. Miyagi, K.C.S. Rao, Mass Spectrom. Rev. 26 (2007) 121–136. I.I. Stewart, T. Thomson, D. Figeys, Rapid Commun. Mass Spectrom. 15 (2001) 2456–2465. O.A. Mirgorodskaya, Y.P. Kozmin, M.I. Titov, R. Korner, C.P. Sonksen, P. Roepstorff, Rapid Commun. Mass Spectrom. 14 (2000) 1226–1232. Y. Oda, K. Huang, F.R. Cross, D. Cowburn, B.T. Chait, PNAS 96 (1999) 6591–6596. ´ P. Bermejo, J.L. Capelo, A. Mota, Y. Madrid, C. Camara, Trac-Trend Anal. Chem. 23 (2004) 654–663. C. Fernandez, C.L. Antonio, R. Conceic¸a˜ o, C. Rial-Otero, J.L. Vaz, Capelo, Anal. Chem. 78 (2006) 2494–2499. ´ ´ D. Lopez-Ferrer, J.L. Capelo, J. Vazquez, J. Proteome Res. 4 (2005) 169–1574. R. Rial-Otero, R.J. Carreira, F.M. Cordeiro, A.J. Moro, M.G. Rivas, L. Fernandez, I. Moura, J.L. Capelo, J. Proteome Res. 6 (2007) 909–912. Data Explorer software user’s guide, section 3, pp. 45–47. X. Yao, A. Freas, J. Ramirez, P.A. Demirev, C. Fenselau, Anal. Chem. 73 (2001) 2836. X. Yao, C. Afonso, C. Fenselau, J. Proteome Res. 2 (2003) 147–152. M.M.R. Talukder, M.M. Zaman, Y. Hayashi, J.C. Wu, T. Kawanishi, Biocatal. Biotransform. 24 (2006) 189–194. E. Bracey, R.A. Stenning, B.E. Brooker, Enzyme Microb. Technol. 22 (1998) 147–151. H.M. Santos, J.L. Capelo, Talanta 73 (2007) 795–802. T.J. Mason, Sonochemistry, Oxford University Press, Oxford, UK, 1999.