Molecular weight determination of plasmid DNA using electrospray ionization mass
spectrometry
Molecular weight determination of plasmid DNA using electrospray ionization mass spectrometry
Xueheng
Cheng
1,+
,
David G.
Camp II
2
,
Qinyuan
Wu
1
,
Ray
Bakhtiar1,
[sect]
,
David L.
Springer
3
,
Brendt J.
Morris3,
[para]
,
James E.
Bruce
1
,
Gordon A.
Anderson
1
,
Charles G.
Edmonds
1,3
and
Richard D.
Smith
1,
*
1
Environmental Molecular Sciences Laboratory, Pacific Northwest National
Laboratory,
Richland
, WA 99352,
USA
,
2
Department of Chemistry, Eastern Oregon State College, La Grande, OR 97850,
USA
and
3
Biology and Chemistry Department, Pacific Northwest Laboratory,
Richland
, WA 99352,
USA
Received June 26, 1995;
Revised November 3, 1995 and Accepted April 11, 1996
ABSTRACT
Ionization and molecular weight (MW) determination of megadalton size plasmid
DNA has been achieved using electrospray ionization (ESI) with Fourier
transform ion cyclotron resonance (FTICR) mass spectrometry. DNA molecules were
shown to remain intact through electrospray ionization by collection on a
specially prepared surface, followed by agarose gel electrophoresis. Individual
highly charged ions of plasmid DNA produced by ESI were trapped in an FTICR
cell for up to several hours and reacted with acetic acid to induce charge
state shifts. Measurements of mass-to-charge ratios for these multiple peaks arising from charge state
shifting give MW measurements of individual ions with an average accuracy of
0.2%. The MW distribution was obtained by measurements for a number of
individual ions from the same sample [plasmid DNA: pGEM-5S MW(cal) = 1.946 MDa], yielding a MW(obs) of 1.95
+- 0.07 MDa for ions clustered in the vicinity of the expected MW.
INTRODUCTION
Mass spectrometry (MS) is the most accurate method for molecular weight (MW)
measurement of relatively small molecules, and is widely used for
identification and in structural determination. The development of matrix-assisted laser desorption ionization (MALDI) (
1
) and electrospray ionization (ESI) (
2
,
3
) now allow routine MS measurements of biopolymers of as large as 100-500 kDa. The ESI technique is especially gentle, generally producing
multiply charged ions of intact molecules. Mass accuracies of ~0.01% can be obtained using conventional mass spectrometers (
2
,
3
) due to the generally limited mass-to-charge ratios (m/z) of the multiply charged ions. Ionization and
detection of molecules in the megadalton range by ESI was reported in early
work (
4
), but no confirmation was possible due to the inability to resolve distinct
charge states as needed for MW determination. Recently the ultra-high resolving power (>10
6
) (
6
-
8
) possible with electrospray ionization-Fourier transform ion cyclotron resonance (ESI-FTICR) (
9
,
10
) has allowed low p.p.m. accuracy mass measurements of biomolecules in the 5-70 kDa range (
6
-
8
). However, MW measurements for larger (>100 kDa) molecules remains difficult
using this approach due to limitations upon the number of charges that can be
usefully contained in the FTICR cell (
9
). The insufficient population for multiplets of charge states and isotope peaks
prevents the assignments of the charge states
(z,
and hence the mass,
m)
since mass measurement for multiply charged ions requires the determination of
m/z
spacings between either different charge state peaks (
11
), isotope peaks (
12
) or peaks due to adducts (having known mass) (
13
).
We have recently developed methods for MW measurements of megadalton molecules
based on the measurements of
individual multiply charged ions using FTICR (
14
-
16
). Our approach involves the manipulation of individual ions (i.e., single ionic
species) trapped in an FTICR cell. The large number of charges for large
molecules provides sufficient sensitivity for individual ion detection (
9
,
15
). Stepwise charge state shifting of individual ions in reaction with selected
reagents produces multiple peaks from the same ion from which charge states and
MW can be calculated (
14
-
16
) as:
Where (
m/z
)
1
and (
m/z
)
2
are the mass-to-charge ratios measured before and after the charge state shifting
reaction, [Delta](
m/z
) = (
m/z
)
2
- (
m/z
)
1
, and A is the mass of the charge-carrying species (A is negative for an adduct and positive when the mass A
is removed from the molecule). It should be noted that in a recent report we
also described an alternative approach for charge determination of very large
individual DNA ions (up to 10
8
Da) based on the measurement of the induced current (and thus the charge) of an
individual ion in the FTICR detection circuitry, but at present this approach
offers lower precision (5-10%) (
17
).
Analysis of an ensemble of ions can, in principle, yield the MW distribution, in
essence, a `mass spectrum' can be constructed one (individual) ion at a time
yielding the MW distribution through measurements of a large number of
individual ions. We have performed such measurements for a 3126 bp plasmid
molecule, pGEM-5S. This plasmid DNA was chosen for these initial studies due to its
unique topomeric characteristics which facilitated purification and allowed
highly specific characterization of the material on the basis of sedimentation
velocity and electrophoretic mobility for comparison with ESI-FTICR results.
MATERIALS AND METHODS
Preparation of plasmid DNA
Plasmid molecules pGEM-5S, a 3126 bp supercoiled closed circular double-stranded plasmid subclone of pGEM-9Z-f(-) which includes the 5S ribosomal RNA gene from
Xenopus borealis
(
18
), were prepared and purified to eliminate contamination by lower mass DNA,
ribonucleic acid and salts. The plasmid had a predicted MW of 1 945 664 Da for
the free acid calculated from the base sequence (
18
). The plasmid was cultured in transformed
Escherichia coli
JM109 and, following alkaline lysis, the supercoiled DNA was separated by CsCl
density gradient centrifugation. Successive 100% ethanol precipitation from
ammonium acetate solution was followed by ultrafiltration with Centricon-100 (Amicon, Danvers, MA). Preparations consisted principally of the
supercoiled species, with small contributions by nicked circular and linearized
plasmid.
Electrophoresis of electrosprayed plasmid DNA
The fate of plasmid upon electrospray ionization was investigated by collection
of samples on conductive surfaces followed by agarose gel electrophoresis
analysis (
5
). Targets consisted of a chemically clean stainless steel foil or aqueous 9 mM
Tris, 9 mM boric acid and 0.2 mM EDTA (TBE) buffered droplet (50 [mu]l) stabilized by 1% agarose gel. The plasmid (10 [mu]g/[mu]l in deionized water) was delivered to the ion source at 0.2-0.5 [mu]l/min, and the targets were placed in atmospheric
pressure at entrance plate of the ESI-MS interface (see below) under otherwise identical conditions for the ESI-MS experiments described next (i.e., in the presence of sheath gas
and counterflow gas, as well as differential pumping behind the plate). After
drying of the target, the plasmid was extracted into 50 [mu]l of TBE buffer. Electrophoretic analysis of the recovered plasmid samples
was in 1% agarose gels buffered in TBE with visualization by ethidium bromide
(0.5 [mu]g/ml).
Mass spectrometry
Initial ESI-MS experiments were performed using a Sciex (Thornhill, ON, Canada) TAGA
6000E triple-quadrupole mass spectrometer with a resolving power (m/[Delta]m) of ~800 and mass range of 1400 (
m/z
). This instrument was equipped with a prototype ESI source having a nozzle-skimmer interface (
19
). For collection of plasmid samples and for acquiring plasmid DNA mass spectra,
the following parameters were used: 10 [mu]M DNA samples in deionized water; sample infusion rate of 0.2 [mu]l/min; spray voltage of -2.5 kV; nozzle voltage of -60 V; skimmer voltage of -150 V. Coaxial flows SF
6
and countercurrent N
2
were employed for the ESI of DNA samples.
FTICR experiments were performed using a modified IonSpec (Irvine, CA) FTICR
with an Oxford 7-Tesla superconducting magnet, and incorporating a modified Analytica
(Branford, CT) ESI source (
7
). Time domain signal acquisition was accomplished using an Omega data system
(version 3.0) and a separate, home-built, extended time domain data acquisition system (
8
,
15
). The plasmid DNA sample was prepared as described above with a final
concentration of 0.2 mg/ml (0.1 [mu]M) in deionized water. The sample was electrospray as above, and ions were
accumulated in the FTICR cell for 1 s at ~10
-4
Torr (nitrogen), collisionally cooled for 15 s and detected using chirp
excitation (sweep rate 35 Hz/[mu]s) and broadband data acquisition (using a sampling rate 200 kHz) for 40 s (
7
). For mass and charge determination, the gas phase reaction of individual DNA
ions with acetic acid was used to induce charge state changes. Mass spectra
were acquired several minutes after the pulsed addition of acetic acid to the
system (which raised the system pressure from 10
-9
to 10
-7
Torr for several seconds) (
15
), which was found to facilitate charge transfer reactions at the desired rate.
RESULTS
Analysis of plasmid DNA on quadrupole mass spectrometer and characterization of
DNA after ESI
At the low mass spectrometric resolution of a quadrupole mass spectrometer, ESI
of the plasmid DNA sample afforded a single broad envelope of ionized species
extending between
m/z
500 and 1400 (the upper
m/z
of the instrument). Maximum relative abundance was found at
m/z
~700 (Fig.
1
) (
5
). This observation suggests the possible production of gas-phase multiply charged molecular anions of the plasmid, but affords no
useful mass measurements since both the mass and charge are unknown. We further
characterized the behavior of the plasmid by electrospraying the nucleic acid
onto targets of stainless steel foil or Tris-EDTA-Borate (TEB) buffer droplets stabilized by 1% agarose gel. After
being electrosprayed onto these targets, the plasmid was extracted with TEB
buffer and examined by 1% agarose gel electrophoresis (Fig.
2
). When electrosprayed onto a dry stainless steel target, the plasmid was
substantially altered, with the production of a wide range of materials of both
higher and lower electrophoretic mobility. However, when the plasmid was
electrosprayed onto a liquid buffer surface, the plasmid remained substantially
intact, with the closed circular form of the nucleic acid opened to the nicked
circular form. No substantial degradation to material of lower mass was
observed. On the basis of these results, it cannot be determined whether the
nicked circular form resulted from the ESI process, the surface deposition, or
subsequent handling prior to electrophoresis. Regardless, these results
indicate that the plasmid DNA is at least substantially intact after being
electrosprayed, and the actual MW measurements using FTICR described below,
confirms the stability of at least some substantial fraction of the double-stranded megadalton nucleic acids through the initial electrospray process
and FTICR measurements.
Analysis of plasmid DNA on FTICR-MS
In our ESI-FTICR experiments of plasmid DNA in the negative ion mode, multiple
deprotonation of phosphodiesters produced ions having a broad
m/z
distribution, similar to the low resolution quadrupole experiments except that
ions appeared at somewhat higher
m/z
in FTICR mass spectra (an observation that may be due to the different ESI-MS interfaces used, or to a greater extent of charge reduction due to the
much longer time before analysis). The FTICR mass spectra consist of a
multitude of narrow peaks with no clear pattern of charge states (or isotopic
peaks), and having a pattern which varies with each subsequent injection of
ions. The evidence that those narrow peaks are due to individual ion
contributions includes the observation of quantized behavior of those ions in
responding to RF excitation, ejection and reaction, and the results of the time-resolved ion correlation (TRIC) experiments, as described previously (
15
). An accurate MW measurement cannot be obtained directly from these spectra,
despite the ultra-high resolution capability of the instrument, since the charge states of
the ions are unknown. By reducing the ion population using recently described
techniques (
14
-
16
), these individual ions can be clearly identified (Fig.
3
). Through repeated application of these ion population trimming techniques, we
typically obtained populations of 1-20 individual ions, suitable for subsequent ion-molecule reaction experiments. Each ion can be observed with a
signal-to-noise ratio of ~20-40:1 due to their large charge (
15
). The presence of trace amounts of gaseous acetic acid induced successive
charge state shifts in
m/z
of individual ions during the course of time-domain acquisitions extending to as long as 40 s. The
m/z
shifting is revealed by performing segmented Fourier transformation of the full
time domain signal and/or extraction of the individual ion time-domain signal (time-resolved ion correlation-TRIC) (
15
), allowing the reactant-product relationship between peaks to be established. Figure
4
shows TRIC stack plots for two individual ions from pGEM-5S DNA, illustrating this temporal behavior. During an acquisition period
of 6 s, the
m/z
of one ion moved twice with approximately equal displacements (Fig.
4
A), while another ion shifted four times (Fig.
4
B). We have observed an individual ion shifting its
m/z
position up to 12 times during the course of a time domain acquisition period
of 27 s (data not shown). Some of the peaks in Figure
4
are also split or broadened due to additional
m/z
shifts of much smaller size. The larger
m/z
shifts evident in Figure
4
are identified as charge state changes induced by either reaction with acetic
acid (equation
1
) or dissociative proton transfer of a DNA-acetic acid complex formed through otherwise non-reactive collisions (equation
2
). Similar proton transfer reactions of multiply protonated biomolecular ions
have been studied previously (
16
,
20
). The smaller
m/z
shifts are attributed to addition/elimination of neutral molecules (e.g.,
acetic acid) (equation
3
). The likely origin of the small
m/z
shifts can be tentatively assigned based on the values of the charge state
shifts which give the approximate charge number (
z
) of the peaks, which in turn allow calculation of the masses of the neutral
species involved in the small
m/z
shifts.
M
n-
+ CH
3
COOH
->
M
(n-1)-
+ CH
3
COO
-
1
M
n-
(CH
3
COOH)
->
M
(n-1)-
+ CH
3
COO
-
2
M
n-
+ CH
3
COOH
<->
M
n-
(CH
3
COOH)
3
The small regularly-spaced
m/z
shifts were not observed when nitrogen gas was added instead of acetic acid. We
also observed that the reactivity of individual DNA ions changes with their
m/z
value, i.e., ions with higher
m/z
are generally less reactive toward charge state shifting. This
m/z
dependent reactivity is qualitatively ascribed to differences in the
contribution of Coulombic (repulsive) forces to the reaction energetics for
multiply charged ions of different
m/z
(
16
,
20
).
DISCUSSION
ACKNOWLEDGEMENTS
We thank Drs B. D. Thrall, S. A. Hofstadler, H. R. Udseth, A. L. Rockwood and M.
G. Sherman for helpful discussions and technical assistance, S. Van Orden for
calculating the isotope distribution for the plasmid DNA and the US Department
of Energy Human Genome Program through the Director, Office of Health and
Environmental Research and through PNNL internal Exploratory Research of
Pacific Northwest Laboratory under Contract DE-AC06-76RLO 1830. D.G.C. II acknowledges partial support from Eastern
Oregon State College. Pacific Northwest National Laboratory is operated by
Battelle Memorial Institute.
REFERENCES
1 Karas, M. and Hillenkamp, F. (1988) Anal. Chem.60, 2299-2301.MEDLINE Abstract
2 Fenn, J. B., Mann, M., Meng, C. K., Wong, S. F. and Whitehouse, C. M. (1989) Science246, 64-71.MEDLINE Abstract
3 Smith, R. D., Loo, J. A., Loo, R. R. O., Busman, M. and Udseth, H. R. (1991) Mass Spectrom. Rev.10, 359-451.
4 Nohmi, T. and Fenn, J. B. (1992) J. Am. Chem. Soc.114, 3241-3246.
5 Edmonds, C. G., Springer, D. L., Morris, B. J., Thrall, B. D. and Camp, D. G. (1993) Proceedings of the 41st ASMS Conference on Mass Spectrometry and Allied Topics 272a-272b.
6 Beu, S. C., Senko, M. W., Quinn, J. P., Wampler, F. M. and McLafferty, F. W. (1993) J. Am. Soc. Mass Spectrom.4, 557-565.
7 Winger, B. E., Hofstadler, S. A., Bruce, J. E., Udseth, H. R. and Smith, R. D. (1993) J. Am. Soc. Mass Spectrom. 4, 566-577.
8 Bruce, J. E., Anderson, G. A., Hofstadler, S. A., Winger, B. E. and Smith, R. D. (1993) Rapid Commun. Mass Spectrom.7, 700-703.
9 Marshall, A. G. and Grosshans, P. B. (1991) Anal. Chem.63, 215A-229A.
10 Henry, K. D., Williams, E. R., Wang, B. H., McLafferty, F. W., Shabanowitz, J. and Hunt, D. F. (1989) Proc. Natl Acad. Sci. USA86, 9075-9078.MEDLINE Abstract
11 Mann, M., Meng, C. K. and Fenn, J. B. (1989) Anal. Chem.61, 1702-1708.
12 Henry, K. D. and McLafferty, F. W. (1990) Org. Mass Spectrom.25, 490-492.
13 McLuckey, S. A., Glish, G. L. and Van Berkel, G. J. (1991) Anal. Chem.63, 1971-1978.MEDLINE Abstract
14 Smith, R. D., Cheng, X., Bruce, J. B., Hofstadler, S. A. and Anderson, G. A. (1994) Nature369, 137-139.
15 Bruce, J. E., Cheng, X., Bakhtiar, R., Wu, Q., Hofstadler, S. A., Anderson, G. A. and Smith, R. D. (1994) J. Am. Chem. Soc.116, 7839-7847.
16 Cheng, X. H., Bakhtiar, R., Van Orden, S. and Smith, R. D. (1994) Anal. Chem.66, 2084-2087.MEDLINE Abstract
17 Chen, R., Cheng, X., Mitchell, D. W., Hofstadler, S. A., Wu, Q., Rockwood, A. L., Sherman, M. G. and Smith, R. D. (1995) Anal. Chem.67, 1159-1163.
18 Thrall, B. D., Mann, D. B., Smerdon, M. J. and Springer, D. L. (1992) Carcinogenesis 1529.
19 Smith, R. D., Loo, J. A., Edmonds, C. G., Barinaga, C. J. and Udseth, H. R. (1990) Anal. Chem.62, 882-899.MEDLINE Abstract
20 McLuckey, S. A., Van Berkel, G. J. and Glish, G. L. (1990) J. Am. Chem. Soc.112, 5668-5670.
21 McLuckey, S. A. and Habibi-Goudarzi, S. (1993) J. Am. Chem. Soc.115, 12085-12095.
22 Little, D. P., Chorush, R. A., Speir, J. P., Senko, M. W., Kelleher, N. L. and McLafferty, F. W. (1994) J. Am. Chem. Soc.116, 4893-4897.
23 Bruce, J. E., Van Orden, S. L., Anderson, G. A., Hofstadler, S. A., Sherman, M. G., Rockwood, A. L. and Smith, R. D. (1995) J. Mass Spectrom.30, 124-133.
24 Wunschel, D. S, Fox, K. F., Fox, A., Bruce, J. E., Muddiman, D. C. and Smith, R. D. (1996) Rapid Commun. Mass Spectrom.10, 29-35.MEDLINE Abstract
25 Cheng, X., Gale, D. C., Udseth, H. R. and Smith, R. D. (1995) Anal. Chem.67, 586-593.
26 Fuerstenau, S. D., Benner, W. H. (1995) Rapid Commun. Mass Spectrom. 9, 1508-1538.
27 Smith, R. D., Wahl, J. H., Goodlett, D. R. and Hofstadler, S. A. (1993) Anal. Chem.65, 574A-584A.
28 Speir, J. P., Senko, M. W., Little, D. P., Loo, J. A. and McLafferty, F. M. (1995) J. Mass Spectrom.30, 39-42.
29 Liu, C., Wu, Q., Harms, A. C. and Smith, R. D. submitted for publication.
*
To whom correspondence should be addressed
Present addresses:
+
Pharmaceutical Discovery Division, Abbott Laboratories, Abbott Park, IL 60064,
USA,
[sect]
Drug Metabolism, Merck Research Laboratory, PO Box 2000, Rahway, NJ 07065, USA
and
[para]
Los Alamos Technical Associates, Inc., 8633 Gage Boulevard, Kennewick, WA 99336,
USA
