1488
Anal. Chem. 1991, 63, 1488-1490
Electrospray Mass Spectrometry: Application of Ion Evaporation Theory to Amino Acids Minoru Sakairi*
Central Research Laboratory, Hitachi Ltd., Kokubunji, Tokyo 185, Japan Alfred L. Yergey
National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland 20892
K. W . Michael Siu, J. C. Yvec Le Blanc, Roger Guevremont, and Shier S. Berman Division of Chemistry, National Research Council of Canada, Ottawa, Ontario, Canada K I A OR9
We describe the result of applylng the Ion evaporath theory to a serles of amlno aclds. The very good correlatlon ( r = 0.98) of the natural logartthms of protonated molecule Intensltles observed by electrospray wlth the dMerence between the hydration free energks of molecules and the gasghase Mndlng free energles of molecules and protons In amlno aclds Is conrlstent wlth the Ion evaporatlon model. It seems that the dMerence In the protonated molecule lntenSnles of amino aclds abtalned by electrospray can be explained by a scheme In whlch protonated molecules In the llquld phase are extracted Into the gas phase after a charged droplet Is formed.
INTRODUCTION Electrospray mass spectrometry has received much attention during the last few years. The unique feature of this method is its ability to generate multiply charged ions from proteins, peptides, and nucleotides and to detect them with high sensitivity by quadrupole mass spectrometers (1-3). Numerous investigations about the ionization mechanism of electrospray have been reported (4-8). One possibility to quantitatively discuss the ionization mechanism is the ion evaporation theory proposed by Iribarne and Thomson (9,101. They reported in this theory that solvated ions were extracted into the gas phase from a charged droplet produced by electrospray. Since they investigated only alkali-metal ions, we have tried to apply the theory to organic compounds. However, the solvation free-energy data of solvated ions of organic compounds are not available. Therefore, we assumed that ions of organic compounds were directly desorbed from a charged droplet to simplify the calculation. In this paper, we report the result of applying the ion evaporation theory to a series of amino acids. EXPERIMENTAL SECTION The detaila of the electrospray maw spectrometric system have already been described (8). The electmspray probe used consisted of an approximately 1-in. stainless steel tube with 100-gm i.d. attached by epoxy glue to a 4-in.x l/ls-in.-o.d. stainless steel tube. The probe tip was positioned approximately 15 mm from a sample aperture and was typically biased to 3 kV with respect to ground. Sample solutions were continuously infused with a syringe pump (Harvard Apparatus, Model 22) at a typical flow rate of 0.5 rL/min. The flow rate fluctuation of the syringe pump was less than 2%. We selected amino acids with nonpolar side chains (alanine, isoleucine, leucine, phenylalanine, tryptophan, and valine) and amino acids with uncharged polar side chains (asparagine,glutamine, glycine, serine, threonine, and tyrosine) from 20 amino acids. The other essential amino acids were passed over for lack of data of reliable solvation free energies or binding free energies
(cysteine, methionine, proline, aspartic acid, glutamic acid, arginine, lysine, and hystidine). The protonated molecule intensities of the amino acids were measured by continuously flowing amino acid solutions (loa M concentration prepared in deionized, distilled water acidified with 0.1% formic acid (pH 2.9)) at a flow rate of 0.5 gL/min. The experiments were performed at room temperature (298 K)under atmospheric pressure (1.01 X 105Pa). Ten replicate analyea were performed. AU amino acids were purchased from Sigma Chemical Co. (St. Louis, MO). CALCULATION As detailed by Iribarne and Thomson (9, IO), ion evaporation may be regarded as a reverse process of placing a gas-phase ion in solution. According to this ion evaporation mechanism, we considered a simple scheme for detecting protonated molecules of amino acids. Here, we assumed that protonated molecules of amino acids were directly desorbed from a charged droplet produced by electrospray, although solvated ions were considered to be extracted in the ion evaporation theory. This is because we cannot estimate the solvation free energies of solvated protonated molecules of amino acids. In this case, the number of detected ions per unit time with a mass spectrometer (I)may be simply expressed as follows: I o : Noblb2TlT2 where No is the number of sample molecules introduced into the ion source of the mass spectrometer per unit time, bl is the efficiency of producing protonated molecules in the liquid phase (the ratio of the concentration of protonated molecules in the liquid phase (C,) to the concentration of sample molecules introduced into the ion source (Co)),b2 is the efficiency of extracting protonated molecules from the liquid phase into the gas phase (the ratio of the concentration of protonated molecules extracted into the gas phase (C,) to the concentration of protonated molecules in the liquid phase (Cl)),Tlis the transmission efficiency of protonated molecules in the ion source, and T2 is the transmission efficiency of protonated molecules in the mass spectrometer. The reaction of protonation of molecules in the liquid phase is expressed as
M
+ SH+
-*
MH+ + S
where M and S represent sample and solvent molecules, respectively. If the liquid-phase binding free energies associated with the reactions M H+ MH+ and S + H+ SH+are expressed as AGBL and AGBs, respectively, bl may be given as follows, using the concentrations of solvated protons (Cp) and solvent molecules (Cs):
+
0003-2700/91/0363-1488802.5010 0 1991 American Chemkal Soclety
-
-
ANALYTICAL CHEMISTRY, VOL. 63, NO. 14, JULY 15, 1991
I
LIQUID PHASE GAS PHASE
+
I
t 4
I
%L
I
M H' I
M'-
-AGs c
I
_ _ _I _
AGBL
-
diff, kJ/mol
-883.8 -880.5 -911 -888.9 -815.9 -865 -898.1 -853.3 -882.6 -815 -887.6 -881.1
9.1 8.1 -24.15 -3.11 8.18 8.37 -25.67 0 -20.98 -21.4 -40.63 -39.41
893.5 889.2 886.3 885.1 884.3 813.4 872.4 853.3 861.6 853.6 841 842.3
AGE,
:
MH'
- - - -M
amino acids
AGBG
isoleucine leucine
tryptophan phenylalanine
1 Figure 1. Energy cycle for calculating the difference between the hydration free energy of a protonated molecule and the IiquMphase binding free energy of a molecule and a proton. AG,, hydration free energy of a protonated molecule; AG&, liquidphase binding free energy of a molecule and a proton; AGm, gas-phase binding free energy of a molecule and a proton: AG-. hyckaikin free energy of a molecule; AG,, hydration free energy of a proton.
where R is the gas constant and Tis the absolute temperature. While using the total solvation free energy of bringing a protonated molecule from infinity into a charged droplet (AG,) (9, IO), b2 may be expressed as
b2 = C2/C1 = exp(-(-AGm)/RT)
(3)
According to the ion evaporation theory, AGm is given by AGTS = AGs AGE (4)
+
where ACs is the solvation free energy of a protonated molecule and AGE is the electrostatic contribution of bringing the protonated molecule from infinity into a charged droplet (9, IO). From eqs 1-4, we can finally obtain I a No(Cp/Cs) exp(-(-AGs AGBL - AGBS AGE)/RT)TIT~( 5 )
+
If Cp and CSare much higher than C,, they can be considered to be constants. The experimental conditions described above satisfy this condition. Further, under identical experimental conditions, it is reasonable to consider that AGBs, AGE, TI, and T2 are constants for all of the protonated molecules of the amino acids. Therefore, if No is constant, In I a AGs - AGBL (6) where In Z is the natural logarithm of I. However, we cannot directly determine the values of AGs - ACBLbecause of the difficulty in accounting for solvation effects on charged species in the liquid phase. Therefore, we replaced the parameter of AGs - AGBLwith the alternative thermodynamic cycle parameters of the gas-phase binding free energy of an amino acid molecule and a proton (AGBG)and the solvation free energies of an amino acid molecule (AGsA) and a proton (AGsp). Since the last three parameters can be determined experimentally or by calculation, Figure 1should lead to a determination of AGs - AGBL AGs - AGBL = -AGBG AGSA + AGSp (7)
+
Since the electrospray is operated at room temperature under atmospheric pressure, we can adopt data at standard conditions. The data of the gas-phase binding free energies of amino acid molecules (M) and protons (H+) associated with the reaction M + H+ MH+ at standard condition (AGBG) can be taken from the compilation of published data (11). The standard solvation free energy of an amino acid molecule (AGsA) was estimated as follows. Since solvation free energies are additive (12, 13), the solvation free energy of an amino acid molecule (RCH2(NH2)COOH)is expressed as (14)
-
-
kJ/mol
contributions of side chains to AGs(RH), kJ/mol
+AGBG
PGSA
AGsA
-
Table I. Gas-Phase Binding Free Energies Associated with the Reaction M H+ MH+ (AGsa), Hydration Free Energies of Amino Acid Molecules [AGs(RH)], and the Values of -AGB0 + AGs(RH)"
- - - -\Gsp - - - - -H'
HI+
AGs(RH)
1489
+ AGs(g1ycine)
(8)
valine
alanine tyrosine glycine
threonine serine asparagine glutamine
"The gas-phase binding free energies and the hydration free energies are taken from refs 11 and 15,respectively. where AGs(RH) and AGs(g1ycine) are the solvation free energies of the compound corresponding to the amino acid side chain and glycine, respectively. In these solvation free energies, AG(RH) can be referred to as the published thermodynamic data (15,161. The compounds corresponding to the side chains of the amino acids that we chose are methane (alanine), isobutane (isoleucine), butane (leucine), methylbenzene (phenylalanine), 3-methylindole (tryptophan), propane (valine), acetamide (asparagine), propionamide (glutamine), methanol (serine), ethanol (threonine), and 4methylphenol (tyrosine). From eqs 6-8 In I a -AGBG + AGs(RH) AGs(glycine) +AGsp (9)
+
is obtained. Since the last two terms are common to all of the amino acids, the difference in the values of -AGBG + AGs(RH) AGs(glycine) AGSPis directly reflected by the difference in the values of -AGBG + AGs(RH). Therefore, eq 9 is finally simplified to In I a -AGBG AGs(RH) (10)
+
+
+
If the argument here is right, a good correlation between In Z and -AGBG + AGs(RH) should be observed.
RESULTS AND DISCUSSION Although the electrospray generally works better with solutions containing methanol, none was used in this study for simplicity in data interpretation. In this case, the solvation free energy simply became the hydration free energy. Table I shows the gas-phase binding free energies associated with the reaction M + H+ MH+ (A&), the hydration free energies of the compounds corresponding to the amino acid side chains, [AGs(RH)],and the values of -AGm + AGs(RH). Figure 2 shows the plot between the values of -AGBG + AGs(RH) and the natural logarithms of the protonated molecule intensities, in arbitrary units, obtained by electrospray (In I) in the amino acids. The correlation coefficient was calculated to be 0.98. In this case, mass discrimination was minimal and was simply ignored, because the peaks of all 12 amino acids occurred over a relatively narrow mass range. The range indicated on each data point represents the standard deviation of the mean of 10 measurements. Judging from the high correlation coefficient, it is apparent that there is a strong relationship between the values of -AG, + AGs(RH) and In Z in the amino acids. This result apparently proves the validity of eq 6. In other words, the protonated molecule intensities of the amino acids by electrospray can be predicted by the values of -A& + AGs(RH). For
-
1490
ANALYTICAL CHEMISTRY, VOL.
63,NO. 14, JULY 15, 1991
free-energy difference in the amino acids. We assumed that protonated molecules were directly desorbed from a charged droplet. However, hydrated protonated molecules must be really desorbed since water attachement to ions results in the hydration free-energy reduction of the ions, as pointed out in alkali-metal ions by Iribarne and Thomson (9,IO). Unfortunately, however, we cannot estimate how many water molecules are attached to the protonated molecules of the amino acids when such ions are desorbed from a charged droplet. Even if we can estimate the number of the water molecules attached, we don’t have a method to calculate the hydration free-energy reduction resulting from water molecule attachment. To the contrary, if we can have such data, we may perfectly succeed in explaining the difference in the protonated molecule intensities of the amino acids. Registry No. Alanine, 56-41-7;isoleucine, 73-32-5;leucine, 61-90-5;phenylalanine, 63-91-2; tryptophan, 73-22-3; valine, 7218-4;asparagine, 70-47-3;glutamine, 56-85-9;glycine, 56-40-6; serine, 56-45-1; threonine, 72-19-5; tyrosine, 60-18-4.
1.4
1.2
1
0.8
0.6
0.4
0.2
0
-0.2
!
18401 [ 8 5 0 ] [8601 [€I701 [8801 [8901 19001
FREE ENEROY (kJ/mol)
+
Flgure 2. Plot of -AG= AGs(RH) versus In I (AGm, gas-phase blndlng free energies of molecules and protons: AG,(RH), hydration free energles of molecules; In I, natural logarithm of the protonated molecule intensltles of the amino aclds) (f = 0.98).
comparison, we also evaluated the correlation coefficient between the values of - A G B ~and In I and the correlation coefficient between the values of AGs(RH) and In I. These were calculated to be 0.12 and 0.78,respectively. These values clearly show that it is important to consider both physical parameters of the gas-phase binding free energy (AC,) and the solvation free energy (AGs(RH)) for interpreting the difference in the protonated molecule intensities of the amino acids. On the other hand, the argument here has a problem that a predicted intensity ratio of protonated molecules between two amino acids with eq 10 is much higher than an observed one. For example, although the predicted intensity ratio of leucine to aspargagine is about 2.5 X lo7 with eq 10,the observed one is only 3.5. Now we have considered that this gap is derived from the overestimation of the hydration
LITERATURE CITED (1) Yamashb, M.; Fenn, J. B. J . Phys. Chem. 1984, 88, 4451. (2) Yamashb, M.; Fenn, J. B. J . Phys. Chem. 1984, 88, 4471. (3) Whitehouse, C. M.; Dreyer, R. N.; Yamashb, M.; Fenn, J. B. Anal. Chem. 1985. 57, 675. (4) Vestal, M. L. Proceedings of the 37th ASMS Conference on Mass Spectrometry and Allled Topics, Mlaml Beach, FL, 1989 p 41. (5) Smith, R. D.; Wseth. H. R.; Loo, J. A.; Edmonde. C. G.; Barlnaga. C. J. Proceedings of the 37th ASMS Conference on Mass Spectrometry and Allied Topics, Miami Beach, FL. 1989; p 391. (6) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Mass Spectrom. Rev. 1990, 9 , 37. (7) Schmelzeisen-Redeker, G.; Biitferlng, L.; Riillgen, F. W. Inf. J . Mess Spectrom. Ion Roc. 1989, 90, 139. (8) Slu, K. W. M.; Gardner, G. J.; Berman, S. S. Org. Mass Spectrom. 1989. 24, 931. (9) Iribarne, J. V.; Thomson, B. A. J . Chem. phys. 1978, 64, 2287. (10) Thomson, B. A.; Iribarne, J. V. J . Chem. Phys. 1979, 7 1 , 4451. (11) Lias, S. G.; Llebman, J. F.; Levin, R. D. J . phys. Chem. Ret. &fa 1984, 13, 695. (12) Kauzmann, W. A&. Proteln Chem. 1959, 14, 1. (13) Nozaki, Y.; Tanford, C. J . W .Chem. 1971, 246, 2211. (14) Ooi, T.; Oobatake, M.; Nemethy, G.; Scheraga. H. A. Roc. Ne#. Acad. Scl. U . S . A . 1987. 84, 3086. (15) Cabani, S.; Gianni, P.; Moliica, V.; Lepori. L. J . solution Chem. 1981, 10, 563. (16) Wolfenden, R.; Anderson, L.; Cullis, P. M.; Southgate, C. C. 8. Bbchemistry, 1981, 20, 849.
RECE~VED for review March 4,1991. Accepted April 18,1991.
