Correlation of Solution and Gas Phase Complexation Assessed by Electrospray Ionization Mass Spectrometry: Application to One-, Two-, and Three-Ring Macrocycles

Article abstract of DOI:10.1021/jo960023p

Electrospray ionization mass spectrometry (ESI-MS) was used to probe multiple cation complexation by C12H25(N18N)(CH2)12(N18N)C12H25, 2, and (18N)CH2C6H4CH2(N18), 3.Complexation of two cations (2Na⁺, 2K⁺, or Na⁺ and K

Full text of DOI:10.1021/jo960023p

J . Org. Chem. 1996, 61, 4693-4697
4693
Cor r ela tion of Solu tion a n d Ga s P h a se Com p lexa tion Assessed by
Electr osp r a y Ion iza tion Ma ss Sp ectr om etr y: Ap p lica tion to On e-,
Tw o-, a n d Th r ee-Rin g Ma cr ocycles
Keshi Wang and George W. Gokel*
Bioorganic Chemistry Program and Department of Molecular Biology & Pharmacology, Washington
University School of Medicine, Campus Box 8103, 660 South Euclid Avenue, St. Louis, Missouri 63110
Received J anuary 2, 1996X
Electrospray ionization mass spectrometry (ESI-MS) was used to probe multiple cation complexation
by C12H25 N18N (CH2)12 N18N (CH2)12 N18N C12H25, 2, and 18N CH2C6H4CH2 N18 , 3. Complex-
ation of two cations (2Na+, 2 K+, or Na+ and K+) by 3 and three cations by 2 (3 Na+, 3 K+, and
mixtures) as well as mixed proton-metallic cation complexes of both were observed. The K+/Na+
cation-binding selectivity of 18-crown-6 was studied by ESI-MS of a methanol solution, and the
selectivity profile was favorably compared with data obtained previously by ion-selective electrode
techniques in the same solvent.
In tr od u ction
multiple cation complexation situation that is difficult
to quantitatively assess in solution.
During the past several years, we have attempted to
develop relatively simple, synthetic organic models that
would exhibit functional properties1 of natural2 cation-
conducting channel systems.3 Our goal in this work has
been to understand how selectivity is achieved in dy-
namic, membrane-spanning, cation-conducting systems.
These model structures4 presumably remain stationary
during cation transport unlike carrier compounds which
move as a complex with the cation from one side of the
membrane to the other. In addition to the obvious
synthetic challenge of such model structures is the
requirement to assess cation binding both in solutions
of known polarity and in lipid bilayer membranes.5
Mass spectrometry has been shown to be a useful
analytical tool for the assessment of cation-ligand
interactions. It is of additional value in the present case
because it is a rapid analytical method and very little
sample is required. Practitioners in this field have been
justifiably wary of adapting the technique to complex-
ation studies because of the concern that the gas phase
does not quantitatively reflect the solution phenomena
under study.6 Their question is not whether the interac-
tions apparent in the gas phase are possible or even
reasonable in solution but whether the spectra accurately
and precisely reflect solution phenomena. Certain as-
sertions to the effect that the relationship is a valid one
have been made in the literature,7 although the disparity
between solution phase equilibria and charge state
distributions for certain peptides has been described.8 We
have recently used the Hammett linear free energy
relationship to demonstrate a correlation between the
solution and gas phase binding behavior of variously
substituted dibenzyl ethers with Na+ and K+ cations.9
In the present report, we use the electrospray ionization
(ESI-MS) method to assess the selectivity of 18-crown-6
binding the cations Na+ and K+. The results are com-
pared with the known selectivities determined by a
variety of methods in the solution phase. We also expand
our previous complexation studies (by fast atom bom-
bardment, FAB-MS) of two-ring complexation10 to include
three-ring macrocycles (by ESI-MS).11
Standard methods for assessing cation complexation
in multiring macrocycles are problematic because mul-
tisite binding equilibria are inherently complicated to
analyze. In channel model systems consisting of two or
multiring macrocycles, the question of whether more than
one ring is involved in complexation at any time is an
important issue as it may bear directly on the mechanism
of transport. In natural channels, flux has been at-
tributed to cation occupancy of adjacent binding sites
with fast displacement of cations by the nearest neighbor
during transport. This transport mechanism requires the
* To whom correspondence should be addressed. Tel.: (314) 362-
9297. FAX: (314) 362-9298 or 7058. E-mail: ggokel@pharmdec.wustl.
educ.
X Abstract published in Advance ACS Abstracts, J une 15, 1996.
(1) Gokel, G. W.; Medina, J . C.; Li, C. Synlett 1991, 677-683.
(2) (a) Stein, W. D. Channels, Carriers, and Pumps; Academic
Press: New York, 1990. (b) Hille, B. Ionic Channels of Excitable
Membranes, 2nd ed.; Sinauer Associates: Sunderland, MA, 1992.
(3) (a) Nakano, A.; Xie, Q.; Mallen, J . V.; Echegoyen, L.; Gokel, G.
W. J . Am. Chem. Soc. 1990, 112, 1287-1289. (b) Murillo, O.; Watanabe,
S.; Nakano, A.; Gokel, G. W. J . Am. Chem. Soc. 1995, 117, 7665-7679.
(4) (a) Xie, Q.; Gokel, G. W.; Hernandez, J . C.; Echegoyen, L. J . Am.
Chem. Soc. 1994, 116, 690-696. (b) Li, Y.; Gokel, G. W.; Hernandez,
J . C.; Echegoyen, L. J . Am. Chem. Soc. 1994, 116, 3087-3096.
(5) (a) Deakyne, C. A.; Meot-ner, M. J . Phys. Chem. 1990, 94, 232.
(b) Takayama, M.; Fukai, T.; Nomura, T.; Yamaguchi, T. Org. Mass
Spectrom. 1991, 26, 655. (c) Zhang, H.; Chu, I.; Leming, S. Dearden,
D. J . Am. Chem. Soc. 1991, 113, 7415. (d) Takahashi, T.; Uchiyama,
A.; Yamada, K., Lynn, B. C.; Gokel, G. W. Tetrahedron Lett. 1992, 33,
3828. (e) Sawada, M.; Okumura, Y.; Shizuma, M.; Takai, Y.; Hidaka,
Y.; Yamada, H.; Tanaka, T.; Kaneda, T.; Hirose, K.; Misumi, S.;
Takahashi, S. J . Am. Chem. Soc. 1993, 115, 7381. (f) Takahashi, T.;
Uchiyama, A.; Yamada, K., Lynn, B. C.; Gokel, G. W. Supramol. Chem.
1993, 2, 177. (g) Brodbelt, J .; Maleknia, S. J . Am. Chem. Soc. 1993,
115, 2837. (h) Gross M. L.; Hu, P. J . Am. Chem. Soc. 1993, 115, 8821.
Resu lts a n d Discu ssion
During the past several years, fast atom bombardment
mass spectrometry has proved useful as a means to
(6) (a) Wang, G.; Cole, R. B. Org. Mass Spectrom. 1994, 29, 419. (b)
Langley, G. J .; Hamilton, D. G.; Grossel, M. C. J . Chem. Soc., Perkin
Trans. 2 1995, 929.
(7) Cheng, Z. L.; Siu, K. W. M.; Guevremont, R.; Berman, S. S. J .
Am. Soc. Mass. Spectrom. 1992, 3, 281.
(8) Wang, G.; Cole, R. B. Org. Mass Spectrom. 1994, 29, 419.
(9) Wang, K.; Han, X.; Gross, R. W.; Gokel, G. W. J . Am. Chem.
Soc. 1995, 117, 7680-7686.
(10) Lynn, B. C.; Mallen, J . V.; Mun˜oz, S.; Kim, M.; Gokel, G. W.
Supramol. Chem. 1993, 1, 195-197.
(11) A preliminary report of this aspect of the work has appeared:
Wang, K.; Han, X.; Gross, R. W.; Gokel, G. W. J . Chem. Soc., Chem.
Commun. 1995, 641.
S0022-3263(96)00023-0 CCC: $12.00 © 1996 American Chemical Society

4694 J . Org. Chem., Vol. 61, No. 14, 1996
Wang and Gokel
assess cation binding.12 The importance of this develop-
ing technique is clear from recent reports of J ohnstone
et al.,13 Dearden et al.,14 Brodbelt et al.,15 Bowers,16 and
others.17 We used this method to probe the binding
properties of organometallic ligands.18 In order to assess
whether corresponding results were obtained in solution
and gas phase studies using the FAB-MS technique, we
surveyed a series of 22 crown and lariat ethers.19 It
proved possible in this case to correlate binding constants
determined in the solution phase with peak intensities
observed by FAB-MS. The relationship of binding
strengths determined in solution and the gas phase by
this method was “semi-quantitative” but the technique
is of obvious value. Based in part upon this success, we
extended studies to the compounds described below.
Com p ou n d Syn th eses. Three compounds were re-
quired for the present study. 18-Crown-6, 1, was pre-
pared by the method described in Organic Syntheses.20
Three-ring macrocycle 2 was prepared by the procedure
recently detailed.3a Two-ring ligand 3 was prepared by
the reaction of R,R′-dibromo-p-xylene with 2 equiv of aza-
18-crown-6 by a modification of the procedure reported
by Sutherland and co-workers.21 The structures of
compounds 1-3 are shown.
nable to the study of molecular complexation because it
involves desorption of the substance(s) under study by
low level energy transfer from accelerated atoms that
bombard the sample probe. Electrospray ionization is
an even “softer” method for desorption, and very little
molecular fragmentation is observed when this technique
is used. Like FAB, ESI can be used to detect noncovalent
interactions, but its “sensitivity” to such complexes is,
in principle at least, greater. Moreover, a volatile solvent
such as methanol or chloroform is amenable to ESI-MS
whereas a less volatile matrix (commonly m-nitrobenzyl
alcohol) is normally used in FAB-MS. The experiments
described here were undertaken on a Finnigan mass
spectrometer having an electrospray interface. Details
of the electrospray ion source have been reported previ-
ously.22
18-Cr ow n -6 Com p lexa tion . One of the most exten-
sively studied cation-complexing agents is 18-crown-6 (1),
which binds a large array of metallic and organic ions.
We used this host as a control compound to explore multi-
crown ether cation binding. Sodium and potassium
cations were used in this “calibration” study. In both
cases, chloride was the counteranion.
The equilibrium constant for the reaction of 18-crown-6
with sodium cation in methanol is 104.35 (log10 KS ) 4.35)
and 106.08 for 18-crown-6 with potassium cation.23 The
binding constant ratio [KS (K+)/KS (Na+)] is 54. Direct
use of the relative mass peak intensities to assess the
18-crown-6 binding selectivity of K+ over Na+ presents
an obvious problem. If the two cations were present in
solution in equimolar concentrations and peak intensity
quantitatively corresponds to cation binding strength, the
ESI-MS peak intensity for [1‚K]+ would be 50-fold that
of [1‚Na]+. The 50:1 peak ratio would introduce consid-
erable error since the low peak intensity measurement
for the latter would be inherently inaccurate. Therefore,
a solution was used that contained a higher concentration
ratio of Na+ over K+.
For the present study, four anhydrous methanol solu-
tions were prepared in which [K+] ) 20 µM and [Na+]
was varied so that [Na+]/[K+] (as the chlorides) ) 5, 10,
15, 20. In all cases, [18-crown-6] ) 2 µM. Electrospray
ionization mass spectra were determined in triplicate for
each solution.24 Under these circumstances, the ratio of
K+ to crown 1 was always 10 ([K+]:[1] ) 10:1) and the
concentration of Na+ was even higher (see Experimental
Section). This experiment was designed so that competi-
tion between Na+ and K+ in the same solution could be
reliably observed.
E lect r osp r a y Ion iza t ion Ma ss Sp ect r om et r y
Meth od . The fast atom bombardment method is ame-
(12) (a) Malhotra, N.; Roepstorff, P.; Hansen, T. K.; Becher, J . J .
Am. Chem. Soc. 1990, 112, 3709. (b) Zhang, H.; Chu, I.-H.; Leming,
S.; Dearden, D. V. J . Am. Chem. Soc. 1991, 113, 7415. (c) Katritzky,
A. R.; Malhotra, N.; Ramanathan, R.; Kemerait, R. C., J r.; Zimmerman,
J . A.; Eyler, J . R. Rapid Commun. Mass Spectrom. 1992, 6, 25. (d)
Sawada, M.; Shizuma, M.; Takai, Y.; Yamada, H.; Kaneda, T.;
J anafusa, T. J . Am. Chem. Soc. 1992, 114, 4405.
(13) (a) J ohnstone, R. A. W.; Lewis, I. A. S. J . Chem. Soc., Chem.
Commun. 1984, 1268. (b) J ohnstone, R. A. W.; Lewis, I. A. S. Int. J .
Mass Spectrom. 1983, 46, 451. (c) J ohnstone, R. A. W.; Lewis, I. A. S.;
Rose, M. E. Tetrahedron 1983, 39, 1597.
(14) Zhang, H.; Chu, I.-H. Leming, S.; Dearden, D. V. J . Am. Chem.
Soc. 1991, 113, 7415.
(15) Liou, C.-C.; Brodbelt, J . S. J . Am. Soc. Mass Spectrom. 1992,
3, 543.
(16) Lee, S.; Wyttenbach, van Helden, G.; Bowers, M. T. J . Am.
Chem. Soc. 1995, 117, 10159.
(17) (a) Bonas, G.; Boss, C.; Vegnon, M. R. Rapid Commun. Mass.
Spectrom. 1988, 2, 88. (b) Paker, D. J . Chem. Soc., Chem. Commun.
1985, 1129. (c) Laali, K.; Lattimer, R. P. J . Org. Chem. 1989, 54, 496.
(18) (a) Medina, J . C.; Goodnow, T. T.; Rojas, M. T.; Atwood, J . L.;
Lynn, B. C.; Kaifer, A. E.; Gokel, G. W. J . Am. Chem. Soc. 1992, 114,
10583-10595. (b) Medina, J . C.; Bert C. Lynn; Rojas, M. T.; Gokel, G.
W.; Kaifer, A. E. Supramol. Chem. 1993, 1, 145-153.
(19) Gokel, G. W.; Trafton, J . E. Cation Binding by Lariat Ethers.
In Cation Binding by Macrocycles; Marcel Dekker: New York, 1990;
p 253.
(20) Gokel, G. W.; Cram, D. J .; Liotta, C. L.; Harris, H. P.; Cook, F.
L. Org. Synth. 1977, 57, 30.
(21) J ohnson M. R., Sutherland, I. O., Newton, R. F. J . Chem. Soc.,
Perkin Trans. 1 1980, 586.
Isotop ic Cor r ection s Requ ir ed for Bin d in g Con -
sta n t Stu d ies. Normal isotopic distributions were as-
sumed to be present in the macrocycles and cations used
in the present study. Isotopic corrections were therefore
applied to the experimental data prior to assessing the
complexation constants. Five peaks are observed in a
typical Na+/K+/18-crown-6 mass spectrum (see Figure 1).
The observed peaks are 1‚Na+ (m/z ) 287.0), its isotope
peak (m/z ) 288.0), 1‚K+ (m/z ) 303.0), 1‚K+ 13C isotope
peak (m/z ) 304.1), and 1‚K+ 41K isotope peak (m/z )
305.0). In order to avoid systematic errors, especially for
small peak intensities, theoretical isotopic corrections
(see Experimental Section) were applied in order to
(22) Weintraub, S. T.; Pinckard, R. N.; Hail, M. Rapid Commun.
Mass Spectrom. 1991, 5, 309.
(23) Gokel, G. W.; Goli, D. M.; Minganti, C.; Echegoyen, L. J . Am.
Chem. Soc. 1983, 105, 6786.
(24) Spectra for the [Na+/K+] ) 5 case were measured once.

Solution and Gas Phase Complexation Assessed by ESI-MS
J . Org. Chem., Vol. 61, No. 14, 1996 4695
C1 is a linear coefficient. If linear, the constant C1 is the
correlation factor between the gas phase spectra and
solution phase. Similarly, I[K ) C2[1‚K+]. Therefore,
+
]
I[Na ]/I[K ] ) [C1/C2][[1‚Na+]/[1‚K+]] or I[Na ]/I[K ] ) [C1/C2]-
[[KS(Na+)/KS(K+)][[Na+]/[K+]]. It is apparent from Figure
2 that when the cation concentration ratio [Na+]/[K+],
increases, the crown-cation complex peak intensity ratio
+
+
+
+
+
+
I[Na ]/I[K ] increases linearly. Thus, the experimental data
support the above assumption. Note that the Y intercept
is close to zero. We infer from this that the above
relationship is not only linear in the experimental range
([Na+]/[K+] ) 5-20) but also in the range of [Na+]/[K+]
) 0-5.
F igu r e 1. Electrospray ionization mass spectrum (ESI-MS)
of 18-crown-6 binding Na+ and K+. [Na+]/[K+] ) 10. [K+] ) 20
µM. [18-Crown-6] ) 2 µM.
On the basis of the above assumption, the selectivity
+
+
ratio K+/Na+ ) KS(K+)/KS(Na+) ) [I[K ]/I[Na ]][[Na ]/[K ]]-
+
+
+
+
[C1/C2]. The calculated results of [I[Na ]/I[K ]] (see Table
1) for all four experiments fall in the range 46 ( 1. Since
KS(K+)/KS(Na+) ≈ 54, the calculated value for [C1/C2] is
1.17, reasonably close to 1. If the gas phase spectra
reflect solution phase phenomena and the peak intensity
represents crown-cation binding strength in solution,
[C1/C2] should equal 1, because in this case, peak inten-
sity depends only on the concentrations of the complexes
and not on their identities. We conclude from the above
data, at least for the case of 18-crown-6 complexing Na+
and K+, that the observation of solution complexation by
the ESI-MS method is valid and nearly quantitative
within the specified concentration range. Moreover, the
relative peak intensity ratios can be used to estimate the
crown ether cation-binding selectivity.
Ta ble 1. P ea k In ten sity Da ta for ESI-MS An a lysis of
18-Cr ow n -6 w ith Sod iu m a n d P ota ssiu m Ca tion sa
obsd intensity values
corrd intensity values
[Na+]/[K+]
I(Na
I(K
I(Na
+
+
+
+
I(K
)
)
)
)
5
10
15
20b
12
100
100
100
100
13
123.2
123.2
123.2
123.2
23 ( 1
36 ( 3
47 ( 2
26 ( 2
41 ( 2
54 ( 2
a
b
[K+] ) 20 µM, [18-crown-6] ) 2 µM. Methanol was used as
solvent. “I” stands for peak intensity of crown-cation complex.
[K+177408n µM, [18-crown-6] ) 1 µM.
b
An a lysis of Mu ltir in g Com p lexa tion . Two sets of
experiments were done with each substrate. In the first,
shown in Table 2 as A, a solution of either 2 or 3 (0.05
mM), KCl (0.05 mM), and NaCl (0.5 mM) was prepared
in 10% v/v CHCl3:CH3OH. In the second set (B), 2 (0.05
mM) , KCl (0.15 mM) and NaCl (0.5 mM) were dissolved
in 10% v/v CHCl3:CH3OH. The masses and assignments
of the observed peaks (compounds 2 and 3) along with
relative intensity data are recorded in Table 2.
The concentrations recorded in columns 1 and 2 in
Table 2 reflect an effort to assess differences in sodium
or potassium selectivities. When the Na+:ligand concen-
tration ratio was 10:1, the base peak in the spectra of 2
or 3 (see column A in Table 2) was the complex between
the ligand and a number of sodium cations equal to the
number of macrorings present. Potassium complexation
was certainly observed, but the spectrum was dominated
by sodium-cation complexes. When the K+ concentra-
tion was increased to equimolar with the number of
cation-binding rings present in 2, an ion having the
composition [2‚3K]3+ was prominent, although the base
peak was [2‚2Na‚K]3+. The latter is expected because of
the differential in cation concentrations.
F igu r e 2. Na+/K+ concentration vs ESI-MS peak intensity.
obtain overall binding intensities. These corrections are
reflected in all tabulated data (Table 1).
There are two issues that should be addressed for the
simpler 18-crown-6 cation-binding case before considering
the situation with two- or three-ring cation binders.
First, it would be useful to establish the relationship
between cation-crown complex peak intensity in the gas
phase and the concentrations prepared in the solutions
used in the experiment. Ideally, this should be a linear
relationship. The second issue concerns the relationship
between the cation selectivities observed for the gas
phase compared to data obtained independently for the
solution phase. The latter goes to the heart of the
question, which is whether and to what extent the mass
spectra reflect solution phenomena. The relationship
between the intensity ratio observed in the mass spec-
trum and the cation concentration ratio in solution is
plotted in Figure 2.
To our knowledge, the data presented herein constitute
the first definitive evidence for triple cation complexation.
In a previous study, we found that FAB-MS could detect
the presence of certain dication complexes of tris(mac-
rocycles).11 In the latter case, the three macrocyclic rings
were separated by 1,6-hexylene spacers. Thus, in the
extended conformation, the rings were closer together
than they are in the present case and electrostatic
repulsion could have been greater. The observation of
[3‚2Na]2+ is particularly notable since the macrorings of
3 are separated by a fairly rigid spacer that is expected
to prevent the sort of ring-ring cooperativity that might
be possible in 2.
Assume that the crown-cation complex peak intensity
(ESI-MS) is proportional to the concentration ratio in
+
+
solution, i.e., I[Na
[1‚Na+] or I[Na ] ) C1[1‚Na+], where
]

4696 J . Org. Chem., Vol. 61, No. 14, 1996
Ta ble 2. Electr osp r a y Ion iza tion Ma ss Sp ectr a for Com p ou n d s 2 a n d 3a
Wang and Gokel
rel int B
ion
m/z
486
rel int Ab
rel int Bc
ion
m/z
rel int A
[2‚3H]3+
12
54
17
98
58
100
9
11
26
28
43
95
25
52
100
83
31
10
17
16
31
62
28
17
26
[2‚3Na‚Cl]2+
[2‚H‚2K‚Cl]2+
[2‚2Na‚K‚Cl]2
[2‚Na‚2K‚Cl]2
[2‚3K‚Cl]2+
[2‚Na]+
780.5
785.1
788.2
795.7
804.2
1478.6
31
<5
21
8
<5
22
20
16
43
36
28
17
[2‚2H‚Na]3+
[2‚2H‚K]3+
[2‚H‚2Na]3+
[2‚H‚Na‚K]3+
[2‚3Na]3+
493.5
498.6
500.7
506.2
508.2
511.5
513.5
518.7
524.2
729
740.1
747.9
750.8
759
767
769.2
777.1
[2‚H‚2K]3+
[2‚2Na‚K]3+
[2‚Na‚2K]3+
[2‚3K]3+
81
17
<5
12
35
9
70
31
8
[3‚2H]2+
315.1
326.2
337.1
345.1
353.1
629
651.6
667.6
709.5
725.4
741.6
<5
5
[3‚H‚Na]2+
[3‚2Na]2+
[3‚Na‚K]2+
[3‚2K]2+
100
25
<5
<5
18
<5
24
9
[2‚2H]2+
[2‚H‚Na]2+
[2‚H‚K]2+
[3‚H]+
[2‚2Na]2+
[3‚Na]+
[2‚Na‚K]2+
[2‚2K]2+
[3‚K]+
[3‚2Na‚Cl]+
[3‚Na‚K‚Cl]+
[3‚2K‚Cl]+
[2‚H‚2Na‚Cl]2+
[2‚H‚Na‚K‚Cl]2+
6
<5
<5
a
b
See Experimental Section for conditions. Column A conditions: [2] ) [3] ) K+ ) 0.05 mM; [Na+] ) 0.5 mM. c Column B conditions:
[2] ) 0.05 mM, [K+]) 0.15 mM, and [Na+] ) 0.5 mM.
Poly(ethylene glycol)s [PEG’s, RO(CH2CH2O)nR, R )
H, CH3) are capable of binding more than one cation
simultaneously. Thus, in methanol solution, a plot of log
KS vs log MW increased linearly.25 Nevertheless, the
ability of the PEG’s to catalyze phase-transfer reactions
was not enhanced by the use of long chains. Instead,
catalysis was correlated to the molarity of PEG chains
available to transfer cations from the polar to the
nonpolar phase. The present case is remarkable in this
context as the macrocycle appears capable of binding
three cations in the gas phase in the absence of a
neutralizing anion or coordinating solvent. A perusal of
Table 2 reveals that no peak observable above the base
line indicates the presence of methanol.
observations presented here. At higher concentrations
of K+ (Table 2, conditions B, [Na+]:[K+] ) 10:3; conditions
A, 10:1), five prominent ions are observed, and all of them
contain K+. Two of the three most intense ions contain
three metallic cations, with the base peak being [2‚
2Na‚K]3+. Moreover, four of these five most intense ions
are triply-charged.
It is also interesting to note that the chloride counter-
ion appears in several ions. It is presumably coordinated
to one or more sodium or potassium cations, although
the ESI-MS technique does not afford structural infor-
mation. It should be noted that chloride is observed only
when more than two metallic cations are present; ions
of the type [(2 or 3)‚cation‚Cl] would be neutral overall
and not detected.
No peak in the spectrum of 3 shows a relative intensity
greater than 25% except [3‚2Na]2+
. Two peaks of a
similar intensity (∼25%, [3‚Na‚K]2+ and [3‚2Na‚Cl]+)
have molecular weights that correspond to the inclusion
of two cations.
Con clu sion
Two goals were accomplished in the present study.
First, the K+/Na+ cation binding selectivity of the simple
complexing agent 18-crown-6 is shown to be similar in
methanol solution whether assessed by ion-selective
electrode techniques or by electrospray mass spectrom-
etry. This is an important validation for the ESI-MS
technique as applied to crown ether chemistry and
suggests that broader applicability to complexation phe-
nomena may be justified. We also demonstrate by use
of this mass spectrometric technique that three cations
may be simultaneously bound by a tris(macrocyclic)
ligand. No previous study confirms this possibility, nor,
to our knowledge, is there any obvious solution technique
that can be used to determine cation-binding constants
for such complex structures. The ESI-MS technique
holds considerable promise for assessing cation interac-
tions with binders that are too structurally complex to
assess by other methods.
Compound 2 possesses three diaza-18-crown-6 sub-
units, and the larger number of macrorings affords a
larger number of possible cation complexes. Under
conditions A, as noted in Table 2, the base peak is [2‚
3Na]3+. A peak observed at m/z 500.7 has the composi-
tion [2‚H‚2Na]3+ and a relative intensity of 98%. The
relative intensity for the corresponding [2‚H‚2K]3+ ion
is <10%. Peaks having relative intensities >50% and
which contain two or three metallic cations are [2‚H‚
Na‚K]3+, [2‚2Na‚K]3+, and [2‚2Na]2+. In addition, an ion
corresponding to [2‚2H‚Na]3+ is observed with a similar,
high relative intensity.
The presence of K+ in several of the major ions
observed in the mass spectra of 2 and 3 suggests that
this ion has a higher binding affinity for the diaza-18-
crown-6 ring system than does Na+ since it is present in
lower concentration than the latter. We have previously
determined the cation complexation constants for N,N′-
di-n-nonyl-4,13-diaza-18-crown-6 with Na+ and K+ in
anhydrous methanol solution. The values for log10 KS
are, respectively, 2.99 and 3.70.26 The corresponding
values for the n-nonyl crown are 2.95 and 3.70. These
solution phase data are consistent with the gas phase
Exp er im en ta l Section
1H-NMR were recorded at 300 or 500 MHz in CDCl3
solvents, and chemical shifts are reported in ppm (δ) downfield
from internal (CH3)4Si. Melting points were determined on a
capillary melting point device in open capillaries and are
uncorrected. Thin layer chromatographic (TLC) analyses were
performed on aluminum oxide 60 F-254 neutral (Type E) with
a 0.2 mm layer thickness or on silica gel 60 F-254 with a 0.2
mm layer thickness. Flash chromatography columns were
packed with silica gel, Merck grade 9385, 230-400 mesh 60
Å.
(25) Gokel, G. W.; Goli, D. M.; Schultz, R. A. J . Org. Chem. 1983,
48, 2837.
(26) Gatto, V. J .; Arnold, K. A.; Viscariello, A. M.; Miller, S. R.;
Morgan, C. R.; Gokel, G. W. J . Org. Chem. 1986, 51, 5373.

Solution and Gas Phase Complexation Assessed by ESI-MS
J . Org. Chem., Vol. 61, No. 14, 1996 4697
All reactions were conducted under dry N2 unless otherwise
noted. All reagents were the best grade commercially avail-
able and were distilled, recrystallized, or used without further
purification, as appropriate. Molecular distillation tempera-
tures refer to the oven temperature of a Kugelrohr apparatus.
Combustion analyses were performed by Atlantic Microlab,
Inc., Atlanta, GA, and are reported as percents.
ber (i.e., cylindrical electrode) and metallized entrance of the
glass capillary were operated at -3.5 kV. A +90 V potential
was applied to the metallized exit of the glass capillary.
A
separate +175 V potential was placed on the tube lens for
acquisition of the positive ions. The temperature of nitrogen
drying gas as it entered the electrospray chamber was set at
100 °C, and the drying gas was held at a constant pressure of
25 psi. A 4-min period of signal averaging was employed for
each spectrum.
18-Cr ow n -6, 1, was prepared according to the Organic
Syntheses procedure.21
N,N′-Bis[12-[N-(N′-d od ecyl)d ia za -18-cr ow n -6]d od ecyl]-
d ia za -18-cr ow n -6, 2, was prepared according to the published
procedure and had analytical data that corresponded to
published values.3a
P r ep a r a tion of Sa m p les. Anhydrous CH3OH was used
for the 18-crown-6-K+/Na+ cation-binding selectivity study.
For studies involving 2 and 3, the solvent was 10% CHCl3 in
CH3OH (v/v). Solutions suitable for study were prepared
(ambient temperature, ∼23 °C) by mixing the appropriate
individual stock solutions to afford the desired final concentra-
tion. The resulting solutions were vortexed and allowed to
stand for g30 min prior to use. Sample solutions were infused
directly into the ESI chamber by use of a syringe pump (flow
rate of 2 µL/min).
Isotop e Cor r ection s. The following data were used in the
isotope correction calculations: abundance, 12C, 100%; 13C, 1.11
(contribution to M + 1: 12 (carbons)1.11 ) 13.32, contribution
to M + 2 : 12 × (12 - 1) × 1.11/200 ) 0.81); 1H, 100; 2H,
0.015 (contribution to M + 1: 24(hydrogens) × 0.015 ) 0.36);
16O, 100; 17O, 0.037 (contribution to M + 1: 6(oxygens) × 0.037
) 0.22); 18O, 0.038 (contribution to M + 2: 6 (oxygens) × 0.038
) 1.22); 39K, 100; 40K, 0.013; 41K, 7.22.
Syn th esis of r′,r′-Bis(a za -18-cr ow n -6)-p-xylen e. R,R′-
Dibromo-p-xylene (264 mg, 1 mmol), 1-aza-18-crown-6 (553 mg,
2.1 mmol), and K2CO3 (1.5 mg, excess) were suspended in
butyronitrile (30 mL). The reactants were heated under reflux
for 4 h. The reaction mixture was cooled and filtered, and the
filtrate was evaporated, giving the crude product which was
purified by column chromatography (Al2O3) to afford a light
yellow oil (500 mg, 80%): 1H-NMR 2.79 (t, 8H, NCH2CH2O),
3.45-3.69 (m, 44H, OCH2 + ArCH2N), 7.25 (s, 4H, aryl-H);
ESI-MS M + Na+ 651.6. The NMR data corresponded to
published values.23
Ma ss Sp ectr a . ESI mass spectra were acquired on a triple
quadrupole tandem mass spectrometer (Finnigan MAT TSQ
700, San J ose, CA) equipped with an electrospray interface
(Analytica of Branford, Branford, CT). The detector of the
instrument is an off-axis continuous dynode electron multiplier
operable from -400 to -3000 V, with
a
variable post-
Ack n ow led gm en t. We thank the NIH for a grant
(GM 36262) that supported this work. We also grate-
fully acknowledge the assistance of Drs. R. W. Gross
and X. Han in the early stages of this work.
acceleration/conversion dynode voltage from -3 to -20 kV for
detection of positive ions. The electrospray ion source and its
functional parts have been described in detail. For all experi-
ments, both the electrospray needle and the skimmer were
operated at ground potential, whereas the electrospray cham-
J O960023P