- Email: [email protected]
Elucidation study of thermal decomposition products of human serum albumin using liquid ionization mass spectrometry
Talanta 57 (2002) 667– 673 www.elsevier.com/locate/talanta
Elucidation study of thermal decomposition products of human serum albumin using liquid ionization mass spectrometry Yukio Yokoyama *, Hisakuni Sato, Masahiko Tsuchiya 1 Department of Analytical Chemistry, Faculty of Engineering, Yokohama National Uni6ersity, 79 -5 Tokiwadai, Hodogaya, Yokohama 240 -8501, Japan Received 6 August 2001; received in revised form 13 February 2002; accepted 19 February 2002
Abstract The origin of the intense but unknown peaks at m/z 235 observed in liquid ionization (LI) mass spectra of middle ear effusion and serum was investigated by using related standard compounds and the collision induced dissociation techniques. The ions were observed as the base peaks in mass spectra of the aqueous fractions of middle ear effusion and serum after chloroform extraction and in those of authentic human serum albumin (HSA) too. The ions commonly observed in serous fluids could be estimated as tyrosil– valil interchain immonium ions arising from thermal decomposition of HSA during the measurement. Such thermally stable interchain immonium ions, also observed in some oligopeptides having Val–Tyr sequence as their fragment ions, are likely to be characteristic ions for large protein molecules. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Liquid ionization mass spectrometry; Val–Tyr interchain immonium ion; Human serum albumin
1. Introduction Liquid ionization (LI) mass spectrometry [1,2], which has been referred to as ‘LPI’ lately [3], is a soft ionization method for organic compounds, providing mainly protonated molecules, [M+ H]+, and several characteristic fragment ions under atmospheric pressure. The ionization process * Corresponding author. Tel./fax: +81-45-339-3939. E-mail address: [email protected] (Y. Yokoyama). 1 Emeritus Professor of Yokohama National University, Present address: 4-37-27 Kugayama, Suginami, Tokyo 1680082, Japan.
is based on the Penning ionization with excited argon produced by corona discharge, which can be briefly expressed as M+ Ar* M+ + e− + Ar (M: sample molecule) and the subsequent proton transfer reaction like M+ + M[M+ H]+ + [M− H] (or H+ can come from matrix molecules), depending on proton affinities. The detailed ionization mechanisms are described in the references cited. Since the optimum temperature for desorption of ions depends on the thermal properties of compounds, analytes in a mixture can be detected separately to some extent by programming the sample temperature. Therefore, the LI method has been applied to the
0039-9140/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 2 ) 0 0 0 7 3 - 5
668
Y. Yokoyama et al. / Talanta 57 (2002) 667–673
analysis of complex mixtures such as engine oil [2], coal tar pitches [4], urine [5– 8], sweat [9], and middle ear effusion (MEE) [10]. The method has also been applied to the discrimination of positional isomers of lipid hydroperoxides [11] by observing their characteristic fragment ions. The use of matrices and/or additives through vapor phase is often effective to observe molecular ion species of thermally labile organic compounds such as lipid hydroperoxides [12]. On the other hand, otitis media with effusion (OME), showing secretory fluids retained in the middle ear cavity without pain, often causes hearing disorders, and the definition of the composition of MEE have been required for the pathogenesis of OME. When carrying out the LI analyses of MEE [10], in those days, the significant peaks at m/z 235 have been observed in the LI mass spectra acquired at higher sample temperature (ca. 200 °C). Besides, the same mass ions have also been observed in an intact blood serum measured as a reference of MEE. In our previous work [10], molecular related ions corresponding to phospholipids, such as phosphatidylcholines and sphingomyelines, and cholesterol have been detected in MEE samples, which have then proved to be secreted from blood sera of the patients. Thus, we estimated that the mass 235 ions also originated from some common constituents of middle ear effusion and serum. It is known that the serum albumin is one of main constituents of MEE. When an authentic human serum albumin (HSA) was measured by the LI method under the same conditions, the intense peaks at m/z 235 were observed in their mass spectra at the sample temperature of around 200 °C. Since a brown-colored residue was left on the sample holder after measurement, we concluded that the ion was produced by thermal decomposition of serum albumin. From the analytical mass spectrometry viewpoint, such thermally stable unknown ions attracted our interest for finding their origins. This paper describes an LI mass spectrometry technique for the structural elucidation of the mass 235 ions, and a possible chemical structure of the ions is also proposed.
2. Experimental
2.1. Materials HSA, bovine serum albumin (BSA), and angiotensin II (Asp–Arg –Val –Tyr –Ile–His –Pro – Phe, FW: 1046) were purchased from Wako (Osaka, Japan). Angiotensin I (Asp–Arg –Val – Tyr –Ile–His –Pro –Phe –His –Leu, FW: 1296), glucagon (His – Ser – Gln – Gry – Thr – Phe – Thr – Ser – Asp – Tyr – Ser – Lys – Tyr – Leu – Asp – Ser – Arg – Arg – Ala – Gln – Asp – Phe – Val – Gln – Trp – Leu –Met –Asn –Thr), and a peptide b–Ala – Trp – Met –Asp –Phe –amide were purchased from Sigma (St. Louis, MO). Other chemicals of guaranteed grade were purchased from Wako. An approximately 3.5% (w/w) aqueous solutions of HSA was prepared for the LI measurements. Approximately 1% (w/w) aqueous solution of each peptide was also prepared. The MEE samples from patients with OME were furnished by Sanraku Hospital (Tokyo, Japan) and human blood sera were obtained from several voluntary students, which were stored at −30 °C before use.
2.2. Sample pretreatment A 200-ml aliquot of the intact serum was transferred into a 1.5-ml micro test tube (Eppendorf, Germany) and diluted with a 200 ml of phosphate buffer (pH 7.4). A 400 ml of chloroform was added to the solution, and the mixture was shaken with vortex for 10 min. The homogenate was then centrifuged (TOMY Model MCX-150, Tokyo, Japan) at 8000× g for 10 min. The resultant organic and aqueous layers were collected separately for the measurements. In the case of middle ear effusion, the extraction was carried out in a reduced scale because of the limitation of sample size.
2.3. Mass spectrometry The instruments used were a modified JEOL (Tokyo, Japan) JMS-QH100 quadrupole mass spectrometer [13] and a modified Shimadzu (Kyoto, Japan) QP-1000 quadrupole mass spectrometer; both equipped with the LI ion sources. The
Y. Yokoyama et al. / Talanta 57 (2002) 667–673
ion focusing procedure and the instrumental conditions were basically the same as described in our previous papers [5–12]. A 0.5-ml aliquot of the sample solution was deposited on the sample holder, which was threedimensionally adjustable to the optimum position [9] to the pinhole. After this, a mass spectrum between m/z 10 and 600 was scanned and recorded in 1s and the scan was repeated every 1.5 s for 150 s (total 100 scans), while the sample temperature was continuously increased from 25 to 240 °C by increasing the heater current [9] from 0 to 0.75 A. Using the Shimadzu instrument, collision induced dissociation (so-called in source conevoltage fragmentation) technique can be performed by increasing the potential difference between skimmer-1 and -2 [13] to 30– 40 V, while a soft ionization condition is available by setting the difference to 10– 20 V. In fact, the fragmentation can be controlled by changing the skimmer-1 voltage because the skimmer-2 voltage is always fixed to approximately + 15 V. All LI mass spectra presented here are obtained
669
by averaging 5–20 mass spectra recorded in succession.
3. Results and discussion
3.1. Origin of the mass 235 ions Fig. 1 shows an LI mass spectrum of middle ear effusion in direct measurement at higher sample temperatures above 200 °C corresponding to the scan numbers between 90 and 100. Very intense but unknown peak was observed at m/z 235 as the base peak of the spectrum. The peak at m/z 114 seems to be protonated molecules, [M+ H]+, of creatinine, relatively abundant metabolite in biological fluids. The mass spectral patterns were very reproducible in replicate measurements. In contrast, the lipophilic constituents in middle ear effusion such as cholesterol and phosphatidylcholines were extracted into chloroform, and the LI mass spectra of the fraction exhibited no peaks at m/z 235 during heating at high temperatures (ca. 200 °C). Besides, the LI mass spectra of the
Fig. 1. LI mass spectrum of middle ear effusion obtained by direct measurement, averaging the scan numbers between 90 and 100.
670
Y. Yokoyama et al. / Talanta 57 (2002) 667–673
Fig. 2. LI mass spectrum of HSA, averaging the scan numbers between 80 and 90.
aqueous fraction obtained from middle ear effusion after chloroform extraction gave almost the same mass spectral patterns as shown in Fig. 1. Furthermore, the close similarity was also observed in the LI mass spectra of the aqueous fractions from sera after chloroform extraction. Similarly, the LI mass spectra of a blistery fluid measured as a reference of serum exhibited significantly intense peaks at m/z 235 together with several characteristic peaks at m/z 70, 114, 209, and 259. Such noteworthy similarity in these mass spectra could naturally lead to our estimation that the ions of mass 235 might originate from some hydrophilic species commonly existing in middle ear effusion and serum. Albumin (66.5 kDa) is the most abundant (ca. 40%) protein in serum. Fig. 2 shows an LI mass spectrum of an authentic HSA at the sample temperatures of around 200 °C, also exhibiting the intense peak at m/z 235. The mass spectral pattern of HSA was quite similar to those of the aqueous fractions from middle ear effusion and serum. The results strongly suggested that the ions of mass 235 observed in all serous fluids could
originate from serum albumin. In addition, BSA also exhibited the intense peaks at m/z 235 in their mass spectra, which were almost the same as those of HSA although there were small differences in mass spectral patterns between them. These albumins can provide very specific mass spectral patterns below m/z 600 despite such large molecules. Judging from the facts that the peaks at m/z 235 were observed at the sample temperatures exceeding 200 °C and a brown-colored residue of the sample was left on the needle tip after measurements, the mass 235 ions were produced through thermal decomposition of the large molecules and were concluded to be one of characteristics directly indicating the presence of HSA.
3.2. Identification of the mass 235 ion Since immonium ions (H2N+CHR) [14] are often observed in the LI mass spectra of amino acids as their fragment ions from the protonated molecules, we first assumed that the mass 235 ion was of this type of ion. In that case, the ion should be composed of at least two amino acid
Y. Yokoyama et al. / Talanta 57 (2002) 667–673
residues. There are five combinations of binary amino acid residues fitting mass 235, i.e. Val– Tyr, Tyr –Val, Phe–Asp, Asp – Phe, and Met– Met. Looking up the amino acid sequences of HSA [15,16] and BSA [17,18], we found the Tyr– Val residue at only one position in both protein sequences, that is, common sequence of – Asp – Glu –Thr –Tyr – Val –Pro – Lys – at residue numbers between 494 and 500 for HSA and 492 and 498 for BSA. In addition, Asp– Phe and Phe –Asp residues are also found in the sequences, but are present at three different positions in each protein sequence. The Met– Met residue cannot be seen there. Thus, several oligopeptides including the residue of Val–Tyr, Tyr–Val, Asp– Phe, or Phe –Asp were then examined by LI– MS. LI mass spectra of angiotensin II, measured under the soft ionization conditions, exhibited the intense peaks at m/z 235, 245, and 263, as shown in Fig. 3A, although [M+ H]+ was not observed because of the instrumental limitation on mass range. The desorption temperature, at which the
671
analyte peak was observed intensely, was around 180 °C. The base peak at m/z 235 is very likely to be the interchain immonium ion of Val–Tyr residue. The peaks at m/z 263 and 245 are corresponding to the y2¦ ions produced by the cleavage between His and Pro, and the fragment ions due to the loss of H2O from the y2¦, [y2¦− H2O]+, respectively. The cone-voltage fragmentation (CID) spectrum of angiotensin II exhibited several characteristic peaks corresponding to immonium ions of amino acid residues such as Pro at m/z 70, His at m/z 110, and Phe at m/z 120, as shown in Fig. 3B. Since the mass 235 ion still appeared as the base peak even under the cone-voltage fragmentation conditions, the ion must be very stable. Angiotensin I also showed similar mass spectra exhibiting the intense peaks of Val–Tyr immonium ion at m/z 235 and those corresponding to the y2¦ at m/z 267. In contrast, a bioactive peptide with the sequence of b–Ala –Trp –Met –Asp –Phe-amide and glucagon gave no peaks at m/z 235 in their mass
Fig. 3. LI mass spectra of angiotensin II measured under (A) soft and (B) CID conditions.
672
Y. Yokoyama et al. / Talanta 57 (2002) 667–673
Fig. 4. LI mass spectra of HSA measured under (A) soft and (B) CID conditions.
spectra despite the presence of Asp– Phe residue. Therefore, the interchain immonium ions consisting of Asp –Phe or Phe – Asp are unlikely to be formed under the experimental conditions. Fig. 4A and Fig. 4B show LI mass spectra of HSA measured under the soft and cone-voltage fragmentation conditions, respectively. The significant increase in the peak intensity at m/z 72 was observed in the CID spectrum, corresponding to the immonium ion of Val that possibly originated from the Tyr–Val residue. The peak at m/z 84 is characteristic fragment of Lys and/or Glu that are present in the common sequence of – Asp – Glu – Thr –Tyr – Val – Pro – Lys – . These were consistent with the estimation mentioned above.
cleavage between Val and Pro, which is consistent with the strong tendency for cleavage N-terminal to Pro. The three-dimensional structure of HSA retrieved from the PDB (Protein Data Bank) data base indicated that the amino-acid sequence of interest (–Asp – Glu –Thr –Tyr –Val –Pro –Lys –) is located outside stereographically [19], and also located opposite the dimeric conformation [20] each other. Such local sequence producing thermally stable interchain immonium ions analyzed by LI mass spectrometry may be related to an active site of this protein. LI mass spectrometry may provide specific information for the characterization study of proteins.
4. Conclusion The results strongly suggest that the intense peak at m/z 235 observed in the LI mass spectra of middle ear effusion and serum is of tyrosil– valil interchain immonium ion, as illustrated in Scheme 1, probably arising from the thermal decomposition of albumin during the measurement. The Tyr –Val immonium ions can basically originate from the loss of CO from the corresponding oxonium ions (Tyr– Val – C O+) due to the
Scheme 1. Possible chemical structure of tyrosil – valil interchain immonium ion.
Y. Yokoyama et al. / Talanta 57 (2002) 667–673
References [1] M. Tsuchiya, H. Kuwabara, Anal. Chem. 56 (1984) 14. [2] M. Tsuchiya, H. Kuwabara, K. Musha, Anal. Chem. 58 (1986) 695. [3] M. Tsuchiya, Mass Spectrom. Rev. 17 (1998) 51. [4] Y. Fujioka, M. Tsuchiya, Bunseki Kagaku 37 (1988) 17. [5] Y. Yokoyama, H. Kakinuma, M. Tsuchiya, Mass Spectroscopy (Tokyo) 37 (1989) 17. [6] Y. Yokoyama, H. Kakinuma, M. Tsuchiya, Bunseki Kagaku 39 (1990) 817. [7] Y. Yokoyama, N. Ohmori, M. Tsuchiya, J. Mass Spectrom. Soc. Jpn. 43 (1995) 189. [8] Y. Yokoyama, N. Ohmori, M. Tsuchiya, Anal. Sci. 12 (1996) 665. [9] Y. Yokoyama, M. Aragaki, H. Sato, M. Tsuchiya, Anal. Chim. Acta 246 (1991) 405. [10] Y. Yokoyama, M. Hashimoto, M. Tsuchiya, R. Yabe, Int. J. Mass Spectrom. Ion Processes 111 (1991) 263.
673
[11] Y. Mochida, Y. Yokoyama, T. Kawai, S. Nakamura, M. Tsuchiya, J. Mass Spectrom. Soc. Jpn. 42 (1994) 207. [12] Y. Yokoyama, Y. Mochida, T. Kawai, S. Nakamura, M. Tsuchiya, J. Mass Spectrom. Soc. Jpn. 44 (1996) 483. [13] K. Otsuka, T. Mizuno, K. Azuma, K. Musha, M. Tsuchiya, Bunseki Kagaku 35 (1986) 12. [14] K. Sato, T. Asada, M. Ishihara, F. Kunihiro, Y. Kammei, E. Kubota, C.E. Costello, S.A. Martin, H.A. Scoble, K. Biemann, Anal. Chem. 59 (1987) 1652. [15] P.Q. Behrens, A.M. Spiekerman, R.J. Brown, Fed. Proc. 34 (1975) 591. [16] M.A. Saber, P. Sto¨ ckbauer, L. Mora´ vek, M. Meloun, Collect. Czech. Chem. Commun. 42 (1977) 564. [17] J.R. Brown, Fed. Proc. 34 (1975) 591. [18] R.G. Reed, F.W. Putnam, T. Peters Jr, Biochem. J. 191 (1980) 867. [19] X.M. He, D.C. Carter, Nature 358 (1992) 209. [20] S. Sugio, S. Mochizuki, M. Noda, A. Kashima, Protein Eng. 12 (1999) 439.
Elucidation study of thermal decomposition products of human serum albumin using liquid ionization mass spectrometry Yukio Yokoyama *, Hisakuni Sato, Masahiko Tsuchiya 1 Department of Analytical Chemistry, Faculty of Engineering, Yokohama National Uni6ersity, 79 -5 Tokiwadai, Hodogaya, Yokohama 240 -8501, Japan Received 6 August 2001; received in revised form 13 February 2002; accepted 19 February 2002
Abstract The origin of the intense but unknown peaks at m/z 235 observed in liquid ionization (LI) mass spectra of middle ear effusion and serum was investigated by using related standard compounds and the collision induced dissociation techniques. The ions were observed as the base peaks in mass spectra of the aqueous fractions of middle ear effusion and serum after chloroform extraction and in those of authentic human serum albumin (HSA) too. The ions commonly observed in serous fluids could be estimated as tyrosil– valil interchain immonium ions arising from thermal decomposition of HSA during the measurement. Such thermally stable interchain immonium ions, also observed in some oligopeptides having Val–Tyr sequence as their fragment ions, are likely to be characteristic ions for large protein molecules. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Liquid ionization mass spectrometry; Val–Tyr interchain immonium ion; Human serum albumin
1. Introduction Liquid ionization (LI) mass spectrometry [1,2], which has been referred to as ‘LPI’ lately [3], is a soft ionization method for organic compounds, providing mainly protonated molecules, [M+ H]+, and several characteristic fragment ions under atmospheric pressure. The ionization process * Corresponding author. Tel./fax: +81-45-339-3939. E-mail address: [email protected] (Y. Yokoyama). 1 Emeritus Professor of Yokohama National University, Present address: 4-37-27 Kugayama, Suginami, Tokyo 1680082, Japan.
is based on the Penning ionization with excited argon produced by corona discharge, which can be briefly expressed as M+ Ar* M+ + e− + Ar (M: sample molecule) and the subsequent proton transfer reaction like M+ + M[M+ H]+ + [M− H] (or H+ can come from matrix molecules), depending on proton affinities. The detailed ionization mechanisms are described in the references cited. Since the optimum temperature for desorption of ions depends on the thermal properties of compounds, analytes in a mixture can be detected separately to some extent by programming the sample temperature. Therefore, the LI method has been applied to the
0039-9140/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 2 ) 0 0 0 7 3 - 5
668
Y. Yokoyama et al. / Talanta 57 (2002) 667–673
analysis of complex mixtures such as engine oil [2], coal tar pitches [4], urine [5– 8], sweat [9], and middle ear effusion (MEE) [10]. The method has also been applied to the discrimination of positional isomers of lipid hydroperoxides [11] by observing their characteristic fragment ions. The use of matrices and/or additives through vapor phase is often effective to observe molecular ion species of thermally labile organic compounds such as lipid hydroperoxides [12]. On the other hand, otitis media with effusion (OME), showing secretory fluids retained in the middle ear cavity without pain, often causes hearing disorders, and the definition of the composition of MEE have been required for the pathogenesis of OME. When carrying out the LI analyses of MEE [10], in those days, the significant peaks at m/z 235 have been observed in the LI mass spectra acquired at higher sample temperature (ca. 200 °C). Besides, the same mass ions have also been observed in an intact blood serum measured as a reference of MEE. In our previous work [10], molecular related ions corresponding to phospholipids, such as phosphatidylcholines and sphingomyelines, and cholesterol have been detected in MEE samples, which have then proved to be secreted from blood sera of the patients. Thus, we estimated that the mass 235 ions also originated from some common constituents of middle ear effusion and serum. It is known that the serum albumin is one of main constituents of MEE. When an authentic human serum albumin (HSA) was measured by the LI method under the same conditions, the intense peaks at m/z 235 were observed in their mass spectra at the sample temperature of around 200 °C. Since a brown-colored residue was left on the sample holder after measurement, we concluded that the ion was produced by thermal decomposition of serum albumin. From the analytical mass spectrometry viewpoint, such thermally stable unknown ions attracted our interest for finding their origins. This paper describes an LI mass spectrometry technique for the structural elucidation of the mass 235 ions, and a possible chemical structure of the ions is also proposed.
2. Experimental
2.1. Materials HSA, bovine serum albumin (BSA), and angiotensin II (Asp–Arg –Val –Tyr –Ile–His –Pro – Phe, FW: 1046) were purchased from Wako (Osaka, Japan). Angiotensin I (Asp–Arg –Val – Tyr –Ile–His –Pro –Phe –His –Leu, FW: 1296), glucagon (His – Ser – Gln – Gry – Thr – Phe – Thr – Ser – Asp – Tyr – Ser – Lys – Tyr – Leu – Asp – Ser – Arg – Arg – Ala – Gln – Asp – Phe – Val – Gln – Trp – Leu –Met –Asn –Thr), and a peptide b–Ala – Trp – Met –Asp –Phe –amide were purchased from Sigma (St. Louis, MO). Other chemicals of guaranteed grade were purchased from Wako. An approximately 3.5% (w/w) aqueous solutions of HSA was prepared for the LI measurements. Approximately 1% (w/w) aqueous solution of each peptide was also prepared. The MEE samples from patients with OME were furnished by Sanraku Hospital (Tokyo, Japan) and human blood sera were obtained from several voluntary students, which were stored at −30 °C before use.
2.2. Sample pretreatment A 200-ml aliquot of the intact serum was transferred into a 1.5-ml micro test tube (Eppendorf, Germany) and diluted with a 200 ml of phosphate buffer (pH 7.4). A 400 ml of chloroform was added to the solution, and the mixture was shaken with vortex for 10 min. The homogenate was then centrifuged (TOMY Model MCX-150, Tokyo, Japan) at 8000× g for 10 min. The resultant organic and aqueous layers were collected separately for the measurements. In the case of middle ear effusion, the extraction was carried out in a reduced scale because of the limitation of sample size.
2.3. Mass spectrometry The instruments used were a modified JEOL (Tokyo, Japan) JMS-QH100 quadrupole mass spectrometer [13] and a modified Shimadzu (Kyoto, Japan) QP-1000 quadrupole mass spectrometer; both equipped with the LI ion sources. The
Y. Yokoyama et al. / Talanta 57 (2002) 667–673
ion focusing procedure and the instrumental conditions were basically the same as described in our previous papers [5–12]. A 0.5-ml aliquot of the sample solution was deposited on the sample holder, which was threedimensionally adjustable to the optimum position [9] to the pinhole. After this, a mass spectrum between m/z 10 and 600 was scanned and recorded in 1s and the scan was repeated every 1.5 s for 150 s (total 100 scans), while the sample temperature was continuously increased from 25 to 240 °C by increasing the heater current [9] from 0 to 0.75 A. Using the Shimadzu instrument, collision induced dissociation (so-called in source conevoltage fragmentation) technique can be performed by increasing the potential difference between skimmer-1 and -2 [13] to 30– 40 V, while a soft ionization condition is available by setting the difference to 10– 20 V. In fact, the fragmentation can be controlled by changing the skimmer-1 voltage because the skimmer-2 voltage is always fixed to approximately + 15 V. All LI mass spectra presented here are obtained
669
by averaging 5–20 mass spectra recorded in succession.
3. Results and discussion
3.1. Origin of the mass 235 ions Fig. 1 shows an LI mass spectrum of middle ear effusion in direct measurement at higher sample temperatures above 200 °C corresponding to the scan numbers between 90 and 100. Very intense but unknown peak was observed at m/z 235 as the base peak of the spectrum. The peak at m/z 114 seems to be protonated molecules, [M+ H]+, of creatinine, relatively abundant metabolite in biological fluids. The mass spectral patterns were very reproducible in replicate measurements. In contrast, the lipophilic constituents in middle ear effusion such as cholesterol and phosphatidylcholines were extracted into chloroform, and the LI mass spectra of the fraction exhibited no peaks at m/z 235 during heating at high temperatures (ca. 200 °C). Besides, the LI mass spectra of the
Fig. 1. LI mass spectrum of middle ear effusion obtained by direct measurement, averaging the scan numbers between 90 and 100.
670
Y. Yokoyama et al. / Talanta 57 (2002) 667–673
Fig. 2. LI mass spectrum of HSA, averaging the scan numbers between 80 and 90.
aqueous fraction obtained from middle ear effusion after chloroform extraction gave almost the same mass spectral patterns as shown in Fig. 1. Furthermore, the close similarity was also observed in the LI mass spectra of the aqueous fractions from sera after chloroform extraction. Similarly, the LI mass spectra of a blistery fluid measured as a reference of serum exhibited significantly intense peaks at m/z 235 together with several characteristic peaks at m/z 70, 114, 209, and 259. Such noteworthy similarity in these mass spectra could naturally lead to our estimation that the ions of mass 235 might originate from some hydrophilic species commonly existing in middle ear effusion and serum. Albumin (66.5 kDa) is the most abundant (ca. 40%) protein in serum. Fig. 2 shows an LI mass spectrum of an authentic HSA at the sample temperatures of around 200 °C, also exhibiting the intense peak at m/z 235. The mass spectral pattern of HSA was quite similar to those of the aqueous fractions from middle ear effusion and serum. The results strongly suggested that the ions of mass 235 observed in all serous fluids could
originate from serum albumin. In addition, BSA also exhibited the intense peaks at m/z 235 in their mass spectra, which were almost the same as those of HSA although there were small differences in mass spectral patterns between them. These albumins can provide very specific mass spectral patterns below m/z 600 despite such large molecules. Judging from the facts that the peaks at m/z 235 were observed at the sample temperatures exceeding 200 °C and a brown-colored residue of the sample was left on the needle tip after measurements, the mass 235 ions were produced through thermal decomposition of the large molecules and were concluded to be one of characteristics directly indicating the presence of HSA.
3.2. Identification of the mass 235 ion Since immonium ions (H2N+CHR) [14] are often observed in the LI mass spectra of amino acids as their fragment ions from the protonated molecules, we first assumed that the mass 235 ion was of this type of ion. In that case, the ion should be composed of at least two amino acid
Y. Yokoyama et al. / Talanta 57 (2002) 667–673
residues. There are five combinations of binary amino acid residues fitting mass 235, i.e. Val– Tyr, Tyr –Val, Phe–Asp, Asp – Phe, and Met– Met. Looking up the amino acid sequences of HSA [15,16] and BSA [17,18], we found the Tyr– Val residue at only one position in both protein sequences, that is, common sequence of – Asp – Glu –Thr –Tyr – Val –Pro – Lys – at residue numbers between 494 and 500 for HSA and 492 and 498 for BSA. In addition, Asp– Phe and Phe –Asp residues are also found in the sequences, but are present at three different positions in each protein sequence. The Met– Met residue cannot be seen there. Thus, several oligopeptides including the residue of Val–Tyr, Tyr–Val, Asp– Phe, or Phe –Asp were then examined by LI– MS. LI mass spectra of angiotensin II, measured under the soft ionization conditions, exhibited the intense peaks at m/z 235, 245, and 263, as shown in Fig. 3A, although [M+ H]+ was not observed because of the instrumental limitation on mass range. The desorption temperature, at which the
671
analyte peak was observed intensely, was around 180 °C. The base peak at m/z 235 is very likely to be the interchain immonium ion of Val–Tyr residue. The peaks at m/z 263 and 245 are corresponding to the y2¦ ions produced by the cleavage between His and Pro, and the fragment ions due to the loss of H2O from the y2¦, [y2¦− H2O]+, respectively. The cone-voltage fragmentation (CID) spectrum of angiotensin II exhibited several characteristic peaks corresponding to immonium ions of amino acid residues such as Pro at m/z 70, His at m/z 110, and Phe at m/z 120, as shown in Fig. 3B. Since the mass 235 ion still appeared as the base peak even under the cone-voltage fragmentation conditions, the ion must be very stable. Angiotensin I also showed similar mass spectra exhibiting the intense peaks of Val–Tyr immonium ion at m/z 235 and those corresponding to the y2¦ at m/z 267. In contrast, a bioactive peptide with the sequence of b–Ala –Trp –Met –Asp –Phe-amide and glucagon gave no peaks at m/z 235 in their mass
Fig. 3. LI mass spectra of angiotensin II measured under (A) soft and (B) CID conditions.
672
Y. Yokoyama et al. / Talanta 57 (2002) 667–673
Fig. 4. LI mass spectra of HSA measured under (A) soft and (B) CID conditions.
spectra despite the presence of Asp– Phe residue. Therefore, the interchain immonium ions consisting of Asp –Phe or Phe – Asp are unlikely to be formed under the experimental conditions. Fig. 4A and Fig. 4B show LI mass spectra of HSA measured under the soft and cone-voltage fragmentation conditions, respectively. The significant increase in the peak intensity at m/z 72 was observed in the CID spectrum, corresponding to the immonium ion of Val that possibly originated from the Tyr–Val residue. The peak at m/z 84 is characteristic fragment of Lys and/or Glu that are present in the common sequence of – Asp – Glu – Thr –Tyr – Val – Pro – Lys – . These were consistent with the estimation mentioned above.
cleavage between Val and Pro, which is consistent with the strong tendency for cleavage N-terminal to Pro. The three-dimensional structure of HSA retrieved from the PDB (Protein Data Bank) data base indicated that the amino-acid sequence of interest (–Asp – Glu –Thr –Tyr –Val –Pro –Lys –) is located outside stereographically [19], and also located opposite the dimeric conformation [20] each other. Such local sequence producing thermally stable interchain immonium ions analyzed by LI mass spectrometry may be related to an active site of this protein. LI mass spectrometry may provide specific information for the characterization study of proteins.
4. Conclusion The results strongly suggest that the intense peak at m/z 235 observed in the LI mass spectra of middle ear effusion and serum is of tyrosil– valil interchain immonium ion, as illustrated in Scheme 1, probably arising from the thermal decomposition of albumin during the measurement. The Tyr –Val immonium ions can basically originate from the loss of CO from the corresponding oxonium ions (Tyr– Val – C O+) due to the
Scheme 1. Possible chemical structure of tyrosil – valil interchain immonium ion.
Y. Yokoyama et al. / Talanta 57 (2002) 667–673
References [1] M. Tsuchiya, H. Kuwabara, Anal. Chem. 56 (1984) 14. [2] M. Tsuchiya, H. Kuwabara, K. Musha, Anal. Chem. 58 (1986) 695. [3] M. Tsuchiya, Mass Spectrom. Rev. 17 (1998) 51. [4] Y. Fujioka, M. Tsuchiya, Bunseki Kagaku 37 (1988) 17. [5] Y. Yokoyama, H. Kakinuma, M. Tsuchiya, Mass Spectroscopy (Tokyo) 37 (1989) 17. [6] Y. Yokoyama, H. Kakinuma, M. Tsuchiya, Bunseki Kagaku 39 (1990) 817. [7] Y. Yokoyama, N. Ohmori, M. Tsuchiya, J. Mass Spectrom. Soc. Jpn. 43 (1995) 189. [8] Y. Yokoyama, N. Ohmori, M. Tsuchiya, Anal. Sci. 12 (1996) 665. [9] Y. Yokoyama, M. Aragaki, H. Sato, M. Tsuchiya, Anal. Chim. Acta 246 (1991) 405. [10] Y. Yokoyama, M. Hashimoto, M. Tsuchiya, R. Yabe, Int. J. Mass Spectrom. Ion Processes 111 (1991) 263.
673
[11] Y. Mochida, Y. Yokoyama, T. Kawai, S. Nakamura, M. Tsuchiya, J. Mass Spectrom. Soc. Jpn. 42 (1994) 207. [12] Y. Yokoyama, Y. Mochida, T. Kawai, S. Nakamura, M. Tsuchiya, J. Mass Spectrom. Soc. Jpn. 44 (1996) 483. [13] K. Otsuka, T. Mizuno, K. Azuma, K. Musha, M. Tsuchiya, Bunseki Kagaku 35 (1986) 12. [14] K. Sato, T. Asada, M. Ishihara, F. Kunihiro, Y. Kammei, E. Kubota, C.E. Costello, S.A. Martin, H.A. Scoble, K. Biemann, Anal. Chem. 59 (1987) 1652. [15] P.Q. Behrens, A.M. Spiekerman, R.J. Brown, Fed. Proc. 34 (1975) 591. [16] M.A. Saber, P. Sto¨ ckbauer, L. Mora´ vek, M. Meloun, Collect. Czech. Chem. Commun. 42 (1977) 564. [17] J.R. Brown, Fed. Proc. 34 (1975) 591. [18] R.G. Reed, F.W. Putnam, T. Peters Jr, Biochem. J. 191 (1980) 867. [19] X.M. He, D.C. Carter, Nature 358 (1992) 209. [20] S. Sugio, S. Mochizuki, M. Noda, A. Kashima, Protein Eng. 12 (1999) 439.