RAPID COMMUNICATIONS IN MASS SPECTROMETRY

Rapid Commun. Mass Spectrom. 2010; 24: 6–10

Published online in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/rcm.4352

The use of acetone as a substitute for acetonitrile in

analysis of peptides by liquid chromatography/
electrospray ionization mass spectrometry
Theodore R. Keppel, Martin E. Jacques and David D. Weis*
Department of Chemistry and the Ralph N. Adams Institute for Bioanalytical Chemistry, University of Kansas, 1251 Wescoe Hall Drive, Lawrence,
KS 66045, USA
Received 7 September 2009; Revised 21 October 2009; Accepted 23 October 2009

The recent worldwide shortage of acetonitrile has prompted interest in alternative solvents for liquid
chromatography/mass spectrometry (LC/MS). In this work, acetone was substituted for acetonitrile in
the separation of a peptide mixture by reversed-phase high-performance liquid chromatography (RP-
HPLC) and in the positive electrospray ionization mass spectrometry (ESI-MS) of individual
peptides. On both C12 and C18 stationary phases, the substitution of acetone for acetonitrile as
the organic component of the mobile phase did not alter the gradient elution order of a five-peptide
retention standard, but did increase peak width, shorten retention times, and increase peak tailing.
Positive ESI mass spectra were obtained for angiotensin I, bradykinin, [Leu5]-enkephalin, and
somatostatin 14 dissolved in both acetonitrile/water/formic acid (25%/75%/0.1%) and acetone/
water/formic acid (25%/75%/0.1%). Under optimized ESI-MS conditions, the mass spectral response
of [Leu5]-enkephalin was increased two-fold when the solvent contained acetone. The substitution of
acetone for acetonitrile resulted in only slight changes in the responses of the remaining peptides. A
higher capillary voltage was required for optimum response when acetone was used. Compared
with acetonitrile/water/formic acid (50/50/0.1%), more interfering species below m/z ¼ 140 were found
in the ESI-MS spectra of acetone/water/formic acid (50/50/0.1%). Copyright # 2009 John Wiley &
Sons, Ltd.

The recent shortage of acetonitrile1 has posed a costly represent an obstacle. A comparison of the physical and
problem for many practitioners of liquid chromatography/ chromatographic properties of acetonitrile, methanol, and
mass spectrometry (LC/MS). In reversed-phase separations, acetone was recently published.2 Compared with methanol
the most commonly used solvents for the organic component and acetonitrile, the two major solvents for RP-HPLC,
of the mobile phase are acetonitrile and methanol. Despite acetone has the volatility of methanol and the viscosity of
being more toxic and having a higher cost, acetonitrile has acetonitrile. Acetone also has the highest eluotropic strength
been favored because of its higher reversed-phase eluent on C18 stationary phase. While electrospray ionization (ESI)
strength, lower viscosity, and shorter wavelength UV cut-off. has been well-studied,3–6 the literature is dominated by
In many cases, methanol does not make an acceptable acetonitrile and methanol. With respect to ESI, acetone has a
substitute for acetonitrile. Particularly in the case of reversed- lower surface tension and higher volatility than acetonitrile
phase separations of peptides, acetonitrile is used almost (CRC Handbook of Chemistry and Physics, online edition,
universally. The recent shortage of acetonitrile has generated accessed 29 April 2009), suggesting that ESI would be more
a flurry of technical notes and product advertisements, most efficient with acetone than with acetonitrile. Finally, we note
of which have focused on reductions in flow rates and/or that the price of acetone is comparable to methanol and
column diameters to minimize acetonitrile consumption. roughly one-third that of acetonitrile (Fisher Optima grade,
Acetone has traditionally been overlooked for reversed- US list price7).
phase high-performance liquid chromatography (RP-HPLC) Recently, Fritz et al.2 compared methanol, acetonitrile, and
because its UV cut-off extends out to 330 nm, but for LC/MS acetone for peptide analysis using LC/MS. Tryptic peptides
applications that do not rely on UV detection, this does not from model proteins were analyzed using HPLC coupled to a
linear ion trap with an ESI source. Peptide elution times as
well as protein sequence coverage were compared for the
three organic solvents. In chromatographic separations, Fritz
*Correspondence to: D. D. Weis, Department of Chemistry and the et al. found that, overall, the peptide retention times were
Ralph N. Adams Institute for Bioanalytical Chemistry, Univer- shorter with acetone than with acetonitrile or methanol.
sity of Kansas, 1251 Wescoe Hall Drive, Lawrence, KS 66045,
USA. However, the retention times of individual peptides were not
E-mail: [email protected] directly compared. The authors also found that acetone and

Copyright # 2009 John Wiley & Sons, Ltd.

 Acetone vs. acetonitrile in LC/ESI-MS analysis of peptides 7

acetonitrile separations resulted in comparable sequence 4

(a) 5
coverage for their test proteins and that sequence coverage 3
was poorer with a methanol-based separation. 2
While the work of Fritz et al.2 demonstrated the feasibility
1
of an acetone-based LC/MS analysis of peptides, the effects
of acetone on chromatographic efficiency and ESI were not
0 1 2 3 4 5 6 7 8
explored. In this study, we investigated how acetone affected
peak retention and resolution in the reversed-phase chro- 4
3
matographic separation of a five-peptide retention standard. (b) 5

We also explored the effect of acetone on the positive ESI of 2

four peptides. Finally, we compared the chemical noise
produced by acetone and acetonitrile. 1

0 1 2 3 4 5 6 7 8
EXPERIMENTAL 4
(c) 5
Materials 3
Optima LC/MS grade acetonitrile and water and Optima 2

grade acetone were all obtained from Fisher Scientific (Fair 1

Lawn, NJ, USA). Formic acid (99þ%, in single-use ampoules)
was obtained from Thermo Fisher (Rockford, IL, USA). Fresh
0 1 2 3 4 5 6 7 8
HPLC solvents were prepared daily. Solvent A was 0.1%
formic acid in water and solvent B was either acetonitrile/ 4
5
water/formic acid (90/10/0.1%) or acetone/water/formic (d) 3
acid (90/10/0.1%). 2

The five-peptide retention standard8 (Thermo Fisher,

1
Rockford, IL, USA) consisted of the peptides RGAGGLGLGK-
[NH2], Ac-RGGGGLGLGK-[NH2], Ac-RGAGGLGLGK-
[NH2], Ac-RGVGGLGLGK-[NH2], and Ac-RGVVGLGLGK- 0 1 2 3 4 5 6 7 8
[NH2], where Ac denotes N-terminal acetylation and [NH2] Retention Time (min)
denotes C-terminal amidation. The peptide retention stan-
Figure 1. Total ion chromatograms of the separation of the
dard was prepared by serial dilution with 0.1% formic acid to
peptide retention standard. (a) C12 column with a 5–15% B
approximately 1 ng mL
1. Aliquots were freeze-dried and
acetonitrile/water gradient, (b) C12 column with a 4–14% B
stored at –208C, then reconstituted with 0.1% formic acid to
acetone/water gradient, (c) C18 column with an 8–18% B
1 ng mL
1 as needed.
acetonitrile/water gradient, and (d) C18 column with a 7–18%
Angiotensin I, bradykinin, [Leu5]-enkephalin, and somato-
B acetone/water gradient. The identities of the peaks are 1:
statin 14 were obtained from Anaspec (San Jose, CA, USA) at
RGAGGLGLGK-[NH2], 2: Ac-RGGGGLGLGK-[NH2], 3: Ac-
>95% purity. Working solutions were prepared by serial
RGAGGLGLGK-[NH2], 4: Ac-RGVGGLGLGK-[NH2], and 5:
dilution from frozen stock to a concentration of 1 pmol mL
1
Ac-RGVVGLGLGK-[NH2]. The columns and makeup of
in 25% acetonitrile or acetone, 75% water, 0.1% formic acid.
mobile phases are described in the Experimental section.

Chromatography
High-performance liquid chromatographic separations were sample loop. C12 and C18 reversed-phase columns were
carried out using an Agilent Technologies 1200 series binary compared in this study. The C12 column was a 50  1.0 mm
capillary pump at a flow rate of 50 mL min
1. Samples were Jupiter Proteo, 4 mm particle size, 90 Å pore diameter
injected with a Cheminert microbore six-port injector valve (Phenomenex, Torrance, CA, USA). The C18 column was a
(Valco Instruments Co. Inc., Houston, TX, USA) using a 1 mL 50  1.0 mm ZORBAX 300SB-C18, 3.5 mm particle size, 300 Å

Table 1. Average retention times and peak widths at half maximum (in min) for the five-peptide retention standard separated using
acetonitrile/water or acetone/water gradients. Average values and standard deviations were obtained from three replicates
C12 column C18 column

acetonitrile/water acetone/water acetonitrile/water acetone/water

Peak tr w½ tr w½ tr w½ tr w½

1 2.129  0.008 0.238  0.003 1.763  0.045 0.274  0.025 2.348  0.012 0.115  0.007 1.708  0.017 0.149  0.007
2 3.481  0.055 0.120  0.014 2.679  0.009 0.151  0.013 3.265  0.012 0.086  0.001 2.434  0.021 0.118  0.003
3 3.955  0.061 0.107  0.008 3.159  0.021 0.148  0.013 3.764  0.010 0.082  0.003 2.894  0.022 0.113  0.002
4 5.247  0.029 0.101  0.004 4.624  0.017 0.152  0.008 5.215  0.017 0.080  0.002 4.550  0.015 0.104  0.001
5 6.571  0.035 0.136  0.007 6.414  0.017 0.189  0.003 6.812  0.023 0.104  0.003 6.410  0.008 0.115  0.001

Copyright # 2009 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2010; 24: 6–10
DOI: 10.1002/rcm
8 T. R. Keppel, M. E. Jacques and D. D. Weis

pore diameter (Agilent Technologies, Santa Clara, CA, USA). an 8–18% B gradient of acetonitrile/water (Fig. 1(c)) aligned
Mobile phase gradients were optimized for each column for best with a 7–18% B gradient of acetone/water (Fig. 1(d)). In
both sets of solvents to obtain comparable retention times all cases, the five peptides were well resolved and eluted
and peak separations for the peptide retention standard. between 1.5 and 7 min. More importantly, acetone did not
Retention standard peptides were eluted with a 5 min change the elution order. For both columns, the retention
gradient, with a hold at the gradient endpoint for 1 min, times of all of the peptide retention standards were shorter
and a return to starting conditions over 30 s. when acetone was used (Table 1).
Acetone caused a loss of resolution with both columns. In
Infusions general, the peaks in the acetonitrile-based separations
Solvent mixtures and individual test peptides were directly (Figs. 1(a) and 1(c)) were sharper than those obtained using
infused into the ESI source at a flow rate of 50 mL min
1 acetone (Figs. 1(b) and 1(d)). In addition, we found that the
through a fused-silica capillary (75 mm i.d.; Upchurch
Scientific, Oak Harbor, WA, USA) using a syringe (GT
Series II; SGE Analytical Science, Austin, TX, USA) and a 20

Intensity (x105)
(a) (b) (c)
syringe pump (model 11 Plus; Harvard Apparatus, Hollis- 15
ton, MA). 10
5
Mass spectrometry
The mass spectrometer used was a model 6220 Accurate- 0
Mass Time-of-Flight (Agilent Technologies, Santa Clara, CA, 432 434 436 432 434 436 432 434 436
USA) equipped with an electrospray ionization source m/z m/z m/z
operated in positive ion mode. Spectra were acquired at a
25
rate of 1 spectrum per second, in 2 GHz extended dynamic Intensity (x105) (d) (e) (f)
20
range mode. Nitrogen was used as the drying gas. For
15
peptide retention standard evaluation, the ESI drying gas
10
had a flow rate of 10 L min
1 at 3258C, with nebulizer
5
pressure at 2.4 bar. The fragmentor, skimmer, and capillary
0
voltages were set to 200 V, 50 V, and 3500 V, respectively. For
individual test peptide infusions, ESI parameters were 353 355 357 353 355 357 353 355 357
optimized to maximize target ion peak intensity. m/z m/z m/z
30
Intensity (x105)

(g) (h) (i)

25
RESULTS AND DISCUSSION 20
15
Peptide separation
10
Recently, Fritz et al.2 compared acetone-based and aceto- 5
nitrile-based LC/MS of a tryptic digest of bovine serum 0
albumin (BSA) using identical gradients. Chromatograms
554 558 562 554 558 562 554 558 562
presented by Fritz et al.2 show a more clustered and faster
m/z m/z m/z
elution of peptides when acetone is substituted for
30
Intensity (x105)

acetonitrile on the same gradient, suggesting a decrease in (j) (k) (l)

25
the retention time of more hydrophobic peptides.2 However,
20
individual peaks were not assigned, making it difficult to 15
predict how substitution of acetone would affect chromato- 10
graphic performance. To complement that study, we used a 5
peptide retention standard mix on both a C12 and a C18 0
column to gauge how peak width and retention time would 545 547 549 545 547 549 545 547 549
be affected by substituting acetone for acetonitrile. A 5–15% B m/z m/z m/z
gradient was used with the C12 column to elute the peptide
retention standards using acetonitrile/water (Fig. 1(a)). The Figure 2. Series of ESI mass spectra showing major peak
mobile phase gradients for the other separations were then intensities for (a–c) angiotensin I [Mþ3H]3þ ion, (d–f) [Leu5]-
optimized in an attempt to align the peaks of all five enkephalin [MþH]þ ion, (g–i) bradykinin [Mþ3H]3þ ion, and (j–
standards and reduce the effects of varied retention times on l) somatostatin 14 [Mþ3H]3þ ion. (a, d, g, j) TOF settings were
peak width. The objective was to determine if identical adjusted to maximize signal intensity of peptides dissolved in
chromatographic performance could be obtained with an 25% acetonitrile/75% water/0.1% formic acid directly infused
optimized acetone gradient. However, we were unable to into the electrospray source. (b, e, h, k) Acetone was sub-
identify a linear gradient of acetone/water that gave elution stituted for acetonitrile as the organic solvent and spectra
results exactly like those of acetonitrile/water. A lower were acquired with TOF settings unchanged from their aceto-
percentage of acetone than acetonitrile was required with nitrile-optimized settings. (c, f, i, l) TOF settings were then
both columns. A 4–14% B acetone/water gradient was adjusted to maximize signal intensity for each peptide dis-
required for the C12 column (Fig. 1(b)). For the C18 column, solved in the acetone-containing solvent.
Copyright # 2009 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2010; 24: 6–10
DOI: 10.1002/rcm
 Acetone vs. acetonitrile in LC/ESI-MS analysis of peptides 9

Table 2. Optimized ESI parameters for infusions of each individual peptide dissolved in either acetonitrile/water/formic acid (25/
75/0.1%) or acetone/water/formic acid (25/75/0.1%)
Maximum ion signal
Capillary voltage (V) Fragmentor voltage (V) Gas flow (L min
1) intensity (105)

Peptide Acetonitrile Acetone Acetonitrile Acetone Acetonitrile Acetone Acetonitrile Acetone

angiotensin I 3300 3500 145 160 12 12 14.9 18.9

[Leu5]-enkephalin 3300 3800 180 190 10 10 24.4 22.8
bradykinin 3500 4000 130 110 12 12 11.9 25.2
somatostatin 14 3700 3800 150 140 12 12 17.4 27.9

C18 column yielded sharper peaks with both solvents. The composition on ESI efficiency will vary from peptide to
peak broadening effect of acetone was less pronounced when peptide based on a number of factors, including surface
using the C18 column, as well. Broadening by acetone is most activity.9 Since [Leu5]-enkephalin is the most hydrophobic of
readily apparent for peak 5 since this peak is the best aligned the peptides we analyzed, its surface affinity may be more
across all four chromatograms. Acetone caused a 38% sensitive to solvent polarity. In less polar acetone, higher
increase in the width of peak 5 with the C12 column, but surface activity may lead to more efficient desolvation
only an 11% increase in the width of peak 5 with the C18 during ESI.9 It is also possible that the enhancement is related
column. Finally, tailing was more pronounced when acetone to [Leu5]-enkephalin being singly charged. In summary, we
was used instead of acetonitrile, most obvious in peaks 2 and have found that substitution of acetone for acetonitrile does
3. Overall, substitution of acetone for acetonitrile did not alter not have a detrimental effect on the formation of positive ions
elution order but did decrease retention time and increase by ESI.
peak width and tailing.

Electrospray ionization efficiency Chemical noise

In order to assess the effect of solvent on ionization efficiency, One issue that often arises upon making a change such as
solutions of angiotensin I, bradykinin, [Leu5]-enkephalin and switching solvents (or even solvent vendors) is the
somatostatin 14 were each prepared in acetonitrile/water/ appearance of new spectral interferents. To investigate this
formic acid (25/75/0.1%) and acetone/water/formic acid issue, we infused acetonitrile/water/formic acid (50/50/
(25/75/0.1%) and directly infused into the mass spec- 0.1%) and acetone/water/formic acid (50/50/0.1%) to make
trometer. Figure 2 shows mass spectra obtained from direct a direct comparison in the absence of chemical noise
infusion of the peptides. ESI parameters were initially originating from the peptides or the LC system (Fig. 3).
adjusted to maximize signal intensity using acetonitrile as Comparing spectra in the m/z 50–500 region for these two
the organic solvent (left column of spectra in Fig. 2). Using solvent mixtures shows that acetone had a more complex
these settings, spectra of the peptides in the acetone solvent mixture of ions at higher intensities. However, none of these
were then acquired (middle column of spectra in Fig. 2). interferents was present at levels that would interfere with
Mass spectra with strong signal intensities were obtained peptide analysis at m/z >140. It has been reported that
when using acetone/water even without re-optimizing the acetone is unstable and can undergo base-catalyzed con-
ESI parameters from the acetonitrile/water settings. Finally, densation reactions to produce diacetone (C6H12O2) and
ESI parameters were optimized to maximize signal intensity
for peptides in the acetone solvent (right column of spectra in
Fig. 2). The optimum drying gas temperature was found to be
3508C for all four peptides in both solvents. The acetonitrile-
and acetone-optimized ESI parameters are presented in
Table 2 along with the maximum ion signals observed for the
base peaks of each peptide. For each peptide, a higher
capillary voltage was necessary to optimize signal intensity
when acetone was in the solvent. In addition, a higher
fragmentor voltage was usually necessary to optimize signal
intensity using acetone, with bradykinin being the exception
in this set of peptides. Drying gas flow rate and temperatures
were the same for either solvent. While there was a variation
in signal intensity between the two solvent systems for each
peptide, switching to acetone/water usually resulted in mass
spectra with higher signal intensities. The mass spectrum of
singly charged [Leu5]-enkephalin almost doubled in inten- Figure 3. ESI-MS spectra from infusion of (a) acetonitrile/
sity in optimized acetone/water conditions from optimized water/formic acid (50/50/0.1%) and (b) acetone/water/formic
acetonitrile/water (Figs. 2(g)–2(i)) while for the other acid (50/50/0.1%). Centroid data with an intensity threshold of
peptides the changes were more subtle. The effect of solvent 500 are shown.
Copyright # 2009 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2010; 24: 6–10
DOI: 10.1002/rcm
10 T. R. Keppel, M. E. Jacques and D. D. Weis

triacetone (C9H18O3) and also condensation/elimination Slight losses in chromatographic performance should be
reactions producing mesityl oxide (C6H10O) and pherone anticipated when substituting acetone for acetonitrile as the
(C9H14O).10 At a fragmentor setting of 200 V, typical for organic solvent in reversed-phase LC/MS peptide separ-
peptide analysis, none of the reported condensation products ation analyses, but one should still weigh the cost and time-
were observed with intensities of >103 (i.e., less than 1% of saving benefit it may provide in times of widespread
the vertical scales in Fig. 2). Upon decreasing the fragmentor acetonitrile shortage. Although we have not observed any
voltage to 150 V, an ion with the mass of protonated mesityl deterioration, we caution readers to determine if their LC
oxide was detected above the background noise (results not systems, particularly the piston seals and other polymeric
shown). Even though the bottle of acetone used for these components, will be compatible with acetone.
measurements had been in the lab for over 6 months, we saw
only limited evidence of the reported condensation products.
As a whole, we find that while acetone had more interfering Acknowledgements
ions in the low m/z part of the mass spectrum, it still makes We thank Dr. Ashok Marwah for suggesting that we consider
an acceptable substitute for acetonitrile. using acetone as an alternative organic solvent. This research
was supported by the University of Kansas through the New
Faculty General Research Fund (TRK) and a Seo Research
CONCLUSIONS
Scholarship (MEJ).
An attractive option when looking to reduce costs in the LC/
MS analysis of peptides is the substitution of acetone for
REFERENCES
acetonitrile as the organic solvent. In this work we
investigated how the substitution of acetone for acetonitrile 1. Tullo A. Chem. Eng. News 2008; 86: 27.
2. Fritz R, Ruth W, Kragl U. Rapid Commun. Mass Spectrom.
affected the chromatographic separation and positive ESI of 2009; 23: 2139.
peptides. Compared to acetonitrile, separation of a five- 3. Van Berkel GJ, Kertesz V. Anal. Chem. 2007; 79: 5510.
peptide mixture with an acetone/water gradient preserved 4. Nguyen S, Fenn JB. Proc. Natl. Acad. Sci. USA 2007; 104: 1111.
5. Hogan CJ, Carroll JA, Rohrs HW, Biswas P, Gross ML. Anal.
the elution order but resulted in shorter retention times. Chem. 2008; 81: 369.
However, acetone decreased the separation efficiency 6. Kostiainen R, Kauppila TJ. J. Chromatogr. A 2009; 1216:
resulting in a 15–50% increase in peak width. For most 685.
7. Available: www.fishersci.com (accessed 15 Oct 2009).
peptides, substitution of acetone for acetonitrile slightly 8. Mant CT, Hodges RS. Requirements for peptide standards to
decreased ESI efficiency, but increasing the capillary and monitor column performance and the effect of column
fragmentor voltages reversed this trend, resulting in slightly dimensions, organic modifiers, and temperature in
reverse-phase chromatography. In High Performance Liquid
increased ionization efficiency. While mass spectra gener- Chromatography of Peptides and Proteins: Separation, Analysis,
ated using acetone/water/formic acid solvent showed and Conformation, Mant CT, Hodges RS (eds). CRC Press:
interferences contributing to higher background at m/z Boca Raton, 1991; 289–295.
9. Cech NB, Enke CG. Mass Spectrom. Rev. 2001; 20: 362.
<140, these ions would not overlap peaks from peptide 10. Sadek PC. The HPLC Solvent Guide, (2nd edn). Wiley-
species at higher m/z. Interscience: New York, 2002.

Copyright # 2009 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2010; 24: 6–10
DOI: 10.1002/rcm