Richard D. Smith Steven M. Fields* Joseph A. Loo Charles J. Barinaga Harold R. Udseth Charles G . Edmonts Chemical Methods and Separations Group Chemical Sciences Department Pacific Northwest Laboratory Richland, Washington
709
Capillary isotachophoresis of peptides and proteins
Electrophoresis 1990. 11. 709-717
Capillary isotachophoresis with UV and tandem mass spectrometric detection for peptides and proteins The application of capillary isotachophoresis (CITP) and combined CITP-mass spectrometry (MS) for peptides and proteins is demonstrated. Separation of simple peptide mixtures, as well as enzymatic digets ofproteins, is also reported using CITP with UV detection. The potential utility of CITP for proteins is demonstrated. Initial studies of combined CITP-MS of enzymatic digests is also demonstrated, showing the potential for rapid sequence determination.
1 Introduction The development of capillary electrophoresis (CE) methods provides a basis for the efficient manipulation and separation of subpicomole quantities of polypeptides and proteins. Recent advances in microscale methods, such as the demonstration of the tryptic digestion of low picomole quantities of proteins using the immobilized enzyme in a small diameter packed reaction column[ 1I, are serving to augment further developments. The use of capillary zone electrophoresis (CZE) for separation of proteins[2,3 1 and recent demonstrations of restriction mapping of large deoxyribonucleotides [4 I have propelled potential C E applications into the realm of conventional electrophoresis, while adding the attributes of speed, relatively simple on-line detection, automation, and reduced sample requirements (femtomole to picomole). A literal explosion of ancillary methods for sample manipulation, derivatization, and detection, as well as new methods of obtaining separation selectivity, are being reported. Additionally, other C E formats are attracting increased interest, with the aim of exploiting the unique features of capillary isotachophoresis (CITP)15, 61, capillary isoelectric focusing (CIEF) [ 7 , 81 capillary electrokinetic chromatography (CEC),[9, 101, and, most recently, capillary polyacrylamide gel electrophoresis (CGE)ll11. As a result, there are concomitant and increasing demands upon detector sensitivity and information density. Mass spectrometry (MS) is potentially an ideal detector for CE. At present, CE-MS interfacing methods are based upon either flow (Le., dynamic) fast atom bombardment (FAB) 112-141 or electrospray ionization (ESI) [15-191. Recently, we have demonstrated new interfacing methods based upon ESI that have greatly extended the utility ofCZE-MS by allowing operation over an essentially unlimited range of flow rates and buffer compositions without degrading C Z E separations [17]. These developments have allowed the first online combination of C I T P with MS (CITP-MS) 1201, which provides an attractive compliment to CZE-MS where (among other situations) greater sample sizes are required. These developments have been augmented by the recognition and growing exploitation of the unique features of ESI which include efficient ionization and the production of multiply charged ions from higher molecular weight compounds [2 1-24 I. Correspondence: Dr. R. D. Smith, Chemical Sciences Department, Pacific Northwest Laboratory, Richland, Washington 99352, USA Abbreviations: CE, capillary electrophoresis: CID, collision induced dissociation; CIEF, capillary isoelectric focusing; CZE, capillary zone electrophoresis; CITP, capillary isotachophoresis; ESI, electrospray ionization: MS, mass spectrometry. 0VCH Verlagsgesellachaft mhH, D-6940 Weinhelm, 1990
One goal of our research is the on-line combination of CITPMS and CITP-tandem MS (e.g. MS/MS) for biopolymers. While CITP has been applied extensively to peptides 125-29 I. the combination with MS benefits from some changes in separation conditions compared to those used previously. As a result, changes in leading and trailing buffers, buffer concentrations, and experimental procedures were investigated with the aim of optimizing the CITP-MS combination. The results presented here, while preliminary, show the potential utility of these methods. In particular, the on-line combination of CITP with tandem MS is shown to offer special promise for sequence determination.
2 Materials and methods 2.1 Samples Biochemical samples were purchased from Sigma (St. Louis, MO, USA) and were used without further purification. Tryptic digests utilized an enzyme-to-substrate ratio of 1:12 in a 1 % ammonium bicarbonate solution at 37 OC for 12 h. Synthetic peptides were kindly provided by Dr. Ian Jardine (Finnigan MAT).
2.2 Instrumentation The CITP experiments utilized simple instrumentation constructed at our laboratory, similar to that described previously for C Z E [ 15-171. These experiments utilized 0.6-2 m lengths of 50 or 100 pm i.d. fused silica capillary tubing suspended between two buffer reservoirs (containing the leading and terminating electrolytes). Both untreated and coated fused silica capillaries were investigated; all results reported were coated capillaries with abondedandcrosslinkedfilmofDB-17 which eliminated electroosmotic flow. The columns were typically rinsed with 1 mL of methanol and conditioned for 10 min by application of 20 kV with the capillary filled with the leading electrolyte. The leading electrolyte in most studies was 10 mM ammonium acetate or potassium acetate/S mM ammonium acetate mixtures (pH 4.5-4.8) prepared in doubly distilled deionized water. The trailing electrolyte was generally 10 m M acetic acid. Reagents were degassed daily by sonication and filtered (0.2 pm) before use. Separations involved two steps. First the sample was introduced as a band (10- I00 nL) at the high voltage end (terminating buffer reservoir) of the capillary and voltage was generally manually ramped over -3 min to the finalvoltagenoted. The high voltage was applied
-
-
-
*
Present address: DOW Chemical, P.O. Box 1398, Pittsburgh, CA 94565, USA 0 173-0835/90/0909 0709 $3.50+.25/0
7 10
R. D. Smith et al.
Electrophoresis 1990.11, 109-711
for typically 10-25 min. The second step involved raising the high voltage reservoir (while continuing to apply high voltages) to cause elution of the seDarated bands Dast the Kratos UV-visible detector (detectioi at 210,254 or i S 0 nm). The detector window was made by removing a small section of the polyimide capillary coating 15-20 cm from the capillary terminus. The window slit width was -2 mm.
2.3 CITP-MS For CITP-MS both 50 pm i.d. untreated and 100 pm i.d. DB- 17 coated (bonded and cross-1inked)fused silica capillary columns were used. Useful results were obtained only with coated columns, as described previously. For most experiments the inlet ofthe capillary was biased to the desiredpotential by an electrode immersed in a buffer solution. The column was loaded with the leading electrolyte, and the head of the capillary and the high voltage electrode (the anode for cationic separations) were placed in the sample reservoir. The sample was loaded into the capillary by electromigration or, preferably, by hydrostatic methods (typically by raising the sample reservoir 4 cm for 30 s). When the desired amount of sample had been loaded, the high voltage was interrupted; the electrode and capillary inlet were then transferred to the trailing electrolyte reservoir, and the voltage reapplied at various rates or steps to begin the separation. The instrumentation for CITP-MS developed at our laboratory has been described elsewhere in detail [ 15-17]. The interface design employs a flowing liquid sheath interface which allow the composition and flow rate of the electrosprayed liquid to be controlled independent of the C E buffer [171. The CITP-MS sheath flow rate (10 % water in methanol at 3 pL/min) and cathode voltage were adjusted to form a stable electrospray at the capillary terminus. The electrical contact is also established through the conductive liquid sheath (typically methanol, acetonitrile, acetone or isopropanol, although small fractions of water or acetic acid are generally added for protein analyses 1301). With this arrangement no significant additional mixing (“dead”) volume (< 10 nL) is produced and analyte contact with metal surfaces is avoided. One problem apparent in the present results is a relatively large variation in ESI signal which apparently resulted from bubble formation near the sheath electrode. The elimination of this problem is the subject of continuing effort. For direct ESI-MS experiments without separation, syringe pumps control the flow of analyte solution and liquid sheath at -0.5 pL/min and 3 pL/min, respectively.
-
2.4 ESI source
The ESI source consists of a 50 or 100 pm i.d. fused silica capillary (the CITP capillary) that protrudes 0.2-0.4 mm from a cylindrical stainless steel electrode[ 171. High voltage, generally -1-4 to +6 kV for positive ions or -5 kV for negative ions, is applied to this electrode. The ESI source (capillary) tip is mounted approximately 1.5 cm from the ion sampling nozzle of the ion sampling orifice (nozzle) of the quadrupole mass spectrometer. A 3-6 L/min countercurrent flow of nitrogen gas is introduced between the nozzle and source to aid desolvation of the highly charged electrospray droplets and to minimize any solvent cluster formation during expansion into the vacuum chamber. Ions enter through the 1 mm diameter orifice and are focused efficiently into a 2 nim diameter skimmer orifice directly in front of the radio frequency (refocusing ouadrupole lens. Typically, + 3 50 V to + 1000Vis
applied to the focusinglens and +200V to thenozzle, while the skimmer is at + 65 V [23,3 11.
3 Results and discussion 3.1 Combination of CE and ESI-MS
Our developmental efforts have, from the outset, focused on the combination of C E and ESI-MS. The history of analytical advances in MS has highlighted the special importance of the combination with separation methods having high selectivity and resolving power in conjunction with the high sensitivity and specificity of MS detection. “Real world” samples are invariable mixtures, often very complex mixtures. Any useful analytical method must accommodate contributions from the sample matrix, interfering substances, etc. The dynamic combination of CE, a separation method of high efficiency and flexibility, with ESI-MS is thus particularly advantageous. Electrophoresis forms a family of related techniques including polyacrylarnide gel electrophoresis, isotachophoresis, isoelectric focusing, gel electrofocusing and two-dimensional electrophoresis for very complex mixtures (i.e. extraordinarly high peak capacity). These methods are of fundamental importance in biochemical analysis where different techniques are applied according to the type of sample and the information required. We have previously described the on-line combination of C Z E with ESI-MS [ 15, 161, the first report of any electrophoretic separation technique in dynamic (i.e. on-line) combination with MS. The development was based upon the recognition that both ends of the C Z E capillary did not have to be immersed in buffer reservoirs, as conventionally practiced. The capillary terminus is adapted as an ESI source which accepts the low (from 0 to 1 pL/min) electroosmotic flow and allows ion production at atmospheric pressure from the electrostatically induced nebulization process.
In earlier versions of the CZE/ESI interface the electrospray dispersion and ionization were accomplished from an electrodeposited metal contact established at the end of the C Z E capillary [ 15, 161. Subsequently, we developed aliquid sheath electrode interface from which the solvent composition and flow rate of the electrospray liquid may be controlled independent of the C Z E buffer (which is desirable since high percentage aqueous and high ionic strength buffers useful in C Z E are not well tolerated by ESI) [ 171. Theelectrical contact is established through the conductive sheath, no additional mixing volumes are produced, no significant positive or negative hydrostatic pressure is produced at the outlet of the electrophoretic capillary (i.e. no induced flow) at normal flow rate of the sheath liquid (1-3 pL/min), and analyte contact with metal surfaces is avoided. This interface provides greatly improved performance and is potentially adaptable to other forms of CE. Because C E relies on analyte charge in solution and the ESI process appears to function most effectively for ionic species, the CE/ESI-MS combination is highly complementary. We have reported the analysis of CZE/MS of a mixture of quaternary ammonium compounds [ 151obtaining over 330 000 plates, an order of magnitude better than obtainable by liquid chromatography (LC) in similar time. In this case sample sizes were 300 femtomoles; however, detection limits of I 10 attomoles are obtainable using single ion detection 1 16I.
-
Capillary isotachophoresis of peptides and proteins
Elecfrophorrsis 1990. 1 1 . 709-111
3.2 Potential of CITP
-w Lys-Trp-Lys
Gly-His-Lys
The steady state concentration ofthe analyte anion for CITP separations is determined by the lead anion concentration and the relative mobilities ofthe solute, lead anion and common cation 1251. Accordingly, if the analyte is much more dilute than the lead anion concentration (which is the case in many analytical separations). the analytes will be concentrated as they separate. In a fully developed separation the concentration of each band is similar and the relative abundance for each component is proportional to the length of the band. Thus, CITP offers the potential for higher sample loading (and increased molar sensitivity), high resolution separations and actual concentration (in many cases) of separated sample bands. Figure 1 shows a typical result of a CITP separation of a simple two-component peptide mixture. In such a case the electrophoretic mobilities of the two cotnponents are quitedifferent and a rapid separation is; obtained yielding near ideal bands. Figures 2 and 3 illustrate separations obtained for an enzymatic digestion of substance P and cytochrome c (primarily due to trypsin but complicated by contamination of chymotrypsin - see later discussion in Section 3.3). In the case of cytochronie c at least 16 peptides are expected, reflected by the complex nature of the isotachopherogram in Fig. 3.
UV 210 nm 0-20 uA in 3 min
i
L_
~
0
min
20
Figure I . CITP separation of two tripeptides.
UV 210 nm ramp to 15 kV
'
11'
40
Figure 2. CITP separation of an enzymatic digest of substance P.
We have previously demonstrated the feasibility of the CITP separation of quaternary phosphonium and ammonium salts, amino acids, and catcholamines 1201 and reference peptides I301 with detection by ESI-MS. Detection limits of approximately 1 0 - I ' have ~ been demonstrated for quaternary phosphonium salts and substantial improvements appear feasible. Sample sizes which can be addressed by CITP are much greater (> 100-fold) than CZE. Samples eluting in CITP are ideally flat-topped bands (the length of the analyte band provides information regarding analyte concentration), well suited to MS as the scan speed (ofthe mass spectrometer need not be challenged by the dynamic nature of "sharp" peaks. Most importantly, however, C l T P provides a relatively pure
MS
I
I
min
0
3.3 CITP with detection by ESI-MS
7 11
GLUCAGON
m/z
34
rnin
44
Figure 3. CITP separation of an enzymatic digest of cytochrome c
MW
Figure 4 . ESI-MS spectrum of glucagon (His Ser Gln Gly 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).
7 12
Electrophoresis lYYO,Il, 709-717
R.D. Smith et a[.
analyte band to the ESI source, without the large concentration of supporting electrolyte demanded by CZE. Thus, CITP/MS has the potential of allowing much greater sensitivities (and analyte ion currents) than may be possible by CZE/MS. For our initial attempt at CJTP-MS of enzymatic digests we selected glucagon, a polypeptide of molecular weight 3483. The features of ESI-MS and ESI-MS/MS for polypeptides are illustrated for the intact glucagon molecule in Figs. 4 and 5. Glucagon (lO-%/L) was introduced as a small band by infusion through the CITP capillary. In Fig. 4 the ESI mass spectrum is dominated by multiply protonated molecular ions of glucagon (3 + to 5 + charge states). A small contribution of a molecular dimer, (2M + 7H)7+,was alsoobserved at m/z 996.
MS/MS
Tryptic Digest Fragments of Glucagon
-
T3
HSQGTFTSDYSK~YLDSR;R !AQDFVQWLMNT TI
TZ
T4
Other enzymatic fragments present (chymotryptic)
-
478'
RAQDFVQW~LMNT 1049
-'
Figure 6 . Expected enzymatic cleavage sites for glucagon.
GLUCAGON
M W 3483
b
B274
m/z Figure 5 . ESI-MS/MS spectra of multiply protonated (3+ to 5 + ) charge states of glucagon (see Fig. 4).
Capillary isotachophoresis of peptides and proteins
Electrophoresis 1990,11, 709-7 17
Figure 5 gives the collision-induced dissociation (CID) mass spectra for each of the 3 + to 5 + charge states of glucagon. The major CID fragments are ascribed to the primary ion
sequence and several assignments are noted in Fig. 5 using the generally accepted nomenclature (augmented by superscript for the fragment ion charge state) 130-321. InspectionofFig. 5
Table 1 . \lalor f‘ragmcnt ions produccd from thc multiply cliarpcd parcnt ions o l glucagon t w p t i c iragments _.____~
_.
Amino acid scq ucnce
Pare t i t ion\ yrodricing :
I’aront prod tieing.
.
I
2 2
[An]
1
[Hll]
2 3 4 5
3
3 3 3
3 3 3 -
.~
I2 Scr GI11
II 10
GIy ‘Ilir
9
6 7 8 9 I0
I’hC Tlir Scr A\P TI 1
7 6 4 3
II
Ser
-7
1 yr
5 4 3
1
2 2
(An] (BnI
I 2 3 4
1.c u
A\P Sc r
2 2 7
3
2 2 2 2 2
3 3 3 3 3 3 3
8
5
3
1 I
2
1
3 3 3 3 ~
2 2 2
2 2 I
2 2 2 2 2
-___
2 2 2 2 2
‘4 -
2 2 >
~.
3
2 2 2 2 2 2 2 2
3 3 3 3 3 3 3
2 2 7 2 2 2 2
I 3 3 3 3 3 3 3
?
3 4 5 6 7 8 9 10 I1
A121 Gln
Asp PilC
Val Gill Try Lcu vet
A\n
11 10 9 8
~
.
~
7 6 5 4 3 2
7 13
2 2 2 7
&
2
3 3
.
_
_
_
7 14
R. 11. Smith e/ ul.
shows that (consistent with other results I3 I I) while some of the amino acid sequence can be deduced (primarily near the C or N termini), large portions of the molecule appear inaccessible using the present experimental C I D conditions. An enzymatic (tryptic) digestion of glucagon would be ex pected to produce four tryptic fragments, as shown by the cleavage site predicted in Fig. 6. Figure 7 shows the total ionization current and UV absorbance (210 and 280 nm) isotachopherograms for the CITP/ESI-MS analysis of a mixture of peptides derived from the digestion with trypsin of glucagon. The time scale shown on the axis refers to the begin ning of the separation. Separation development in CITP may be accomplished before detection starts and, as shown here, the actual time for CITP/MS analysis is determined from the time of application of a hydrostatic head to the trailing electrolyte reservoir to cause elution of the focused bands (here at 27 min). The traces show typical bands and step changes at boundaries between bands. The centers of sample bands show no obvious structure. Strong absorbance at 280 nm indicates the presence of tryptophan in the final band eluting. The total reconstructed ion current for MS detection follows the general features of the UV traces, with better characterized bands apparent at the beginning and the end of the separation and alargely structureless region in the middle. Figure 8 further gives selected ion isotachopherogramsfor the tryptic products arginine(Arg+,m/z 175), T1” (m/z 676) and T2’ (m/z 653) for this analysis. Also detected are additional peptide fragments which arise from the action of contaminating chymotrypsin on the predicted tryptic fragments (T4-478)*’ and (T4-1049)’ and on intact glucagon ( m / z 546 and 366). The mass spectrum for the T1 tryptic fragment of glucagon is given in Fig. 9. The ESI mass spectrum of the TI peptide contains intense 3 + and 2 + peaks, while the singly charged molecular ion is observed to have much lower intensity.
3.4 Primary structural information An important aim of our CZE-MS and CITP-MS efforts is to develop methods that will yield primary structural information (i.e. sequence) for polypeptides and small proteins. The ESI-MS method affords unique opportunities in this regard since ionization efficiencies are high and good results can be obtained even for large proteins 123,311. The fact that ESI mass spectra generally consist of only intact multiply charged molecular ions is sometimes cited as a disadvantage of this method since it is claimed that structural information cannot be obtained. However, as we have shown recently, effective dissociation of molecular ions can be induced in the nozzleskimmer region of the ESI interface [ 3 1,33 I. A more powerful approach is tandem MS of molecular ions from several of the major charge states [31,331. The relatively long and stable period of elution of separated bands in CITP should facilitate MS/MS experiments requiring longer integration, signal averaging or more concentrated samples than provided by CZE. While our current data system prevents obtaining MS/MS spectra of more than one charge state from a single separation, future modifications should facilitate this and more flexible methods for selecting multiple peaks for MS/MS study.
klectrophoresis 1990. 11, 109-7 l i
ENZYMATIC DIGEST OF GLUCAGON 11%;
TOTHL
XZI?
l-i
”1 t
I--
Figure 7. Electrospray total ion current isotachopherogram and corresponding UV absorbance(210and 280nm)tracesfortheCITP/MS analysis of a tryptic digest of glucagon. 2 m x 100 prn column, leading buffer 0.0 1 M ammonium pH 4.9; sample concentration 0.0001 M, 25 YO column volume loading; trailing buffer 0.01 M acetic acid. pH 3.3.
-
The C I D tandem mass spectra for the various tryptic fragments (currently obtained in separate experiments due to limitations of our data acquisition system) can be assembled as previously for the smaller polypeptide, melittin 1321. MS/ MS spectraforthevariouschargestates(1 t t o 3 +)fortheTl fragment ofglucagon are shown in Fig. 10. I n Table 1 we give the charge states of various C1D fragments from MS/MS spectra observed for each of the synthetic tryptic fragments. The C I D fragments observed, primarily due to “B” and “Y” mode cleavage of the polypeptides chain [ 31-331, provide a large amount of redundant information regarding the glucagon sequence. As illustrated by Table 1 , major fragment ions are observed which provide information on the entire glucagon sequence.
4 Concluding remarks C E in combination with MS is making increasing contributions to the study of biopolymers. The MS measurement, the determination of mass via the measurement of the ratio of mass to charge, is qualitatively important but (except for very small species) insufficient for full characterization, given a range of possible charge states. Qualitative investigations of complex systems proceed in combined studies with other techniques, eg., nuclear magnetic resonance. crystallography, etc. Of these, MS is frequently the initial experiment by virtue of its relatively high sensitivity. Similari~. Improved separation methods based upon C E will play an important role in such applications. The concurrent developments in both MS and C E are providing complementary tools for manipula
7 15
Capillary isotachophoresis of peptides and proteins
Electrophoresis 1990, 1 I , 709-7 17
-
#,
mlz
175
mlz
546
I ,i----,i# ,A,# #
1 (74;
Figure 8. Selected electrospray ion current isotachopherograms for the analysis of an enzymatic digest with trypsin (and contaminating chymotrypsin) of glucagon.
30
27
MS
- ,
,
,
,
35
,
,
-mi:, -366
0, ,: ,
40
GLUCAGON T R Y P T I C FRAGMENT T 1
709
Capillary isotachophoresis of peptides and proteins
Electrophoresis 1990. 11. 709-717
Capillary isotachophoresis with UV and tandem mass spectrometric detection for peptides and proteins The application of capillary isotachophoresis (CITP) and combined CITP-mass spectrometry (MS) for peptides and proteins is demonstrated. Separation of simple peptide mixtures, as well as enzymatic digets ofproteins, is also reported using CITP with UV detection. The potential utility of CITP for proteins is demonstrated. Initial studies of combined CITP-MS of enzymatic digests is also demonstrated, showing the potential for rapid sequence determination.
1 Introduction The development of capillary electrophoresis (CE) methods provides a basis for the efficient manipulation and separation of subpicomole quantities of polypeptides and proteins. Recent advances in microscale methods, such as the demonstration of the tryptic digestion of low picomole quantities of proteins using the immobilized enzyme in a small diameter packed reaction column[ 1I, are serving to augment further developments. The use of capillary zone electrophoresis (CZE) for separation of proteins[2,3 1 and recent demonstrations of restriction mapping of large deoxyribonucleotides [4 I have propelled potential C E applications into the realm of conventional electrophoresis, while adding the attributes of speed, relatively simple on-line detection, automation, and reduced sample requirements (femtomole to picomole). A literal explosion of ancillary methods for sample manipulation, derivatization, and detection, as well as new methods of obtaining separation selectivity, are being reported. Additionally, other C E formats are attracting increased interest, with the aim of exploiting the unique features of capillary isotachophoresis (CITP)15, 61, capillary isoelectric focusing (CIEF) [ 7 , 81 capillary electrokinetic chromatography (CEC),[9, 101, and, most recently, capillary polyacrylamide gel electrophoresis (CGE)ll11. As a result, there are concomitant and increasing demands upon detector sensitivity and information density. Mass spectrometry (MS) is potentially an ideal detector for CE. At present, CE-MS interfacing methods are based upon either flow (Le., dynamic) fast atom bombardment (FAB) 112-141 or electrospray ionization (ESI) [15-191. Recently, we have demonstrated new interfacing methods based upon ESI that have greatly extended the utility ofCZE-MS by allowing operation over an essentially unlimited range of flow rates and buffer compositions without degrading C Z E separations [17]. These developments have allowed the first online combination of C I T P with MS (CITP-MS) 1201, which provides an attractive compliment to CZE-MS where (among other situations) greater sample sizes are required. These developments have been augmented by the recognition and growing exploitation of the unique features of ESI which include efficient ionization and the production of multiply charged ions from higher molecular weight compounds [2 1-24 I. Correspondence: Dr. R. D. Smith, Chemical Sciences Department, Pacific Northwest Laboratory, Richland, Washington 99352, USA Abbreviations: CE, capillary electrophoresis: CID, collision induced dissociation; CIEF, capillary isoelectric focusing; CZE, capillary zone electrophoresis; CITP, capillary isotachophoresis; ESI, electrospray ionization: MS, mass spectrometry. 0VCH Verlagsgesellachaft mhH, D-6940 Weinhelm, 1990
One goal of our research is the on-line combination of CITPMS and CITP-tandem MS (e.g. MS/MS) for biopolymers. While CITP has been applied extensively to peptides 125-29 I. the combination with MS benefits from some changes in separation conditions compared to those used previously. As a result, changes in leading and trailing buffers, buffer concentrations, and experimental procedures were investigated with the aim of optimizing the CITP-MS combination. The results presented here, while preliminary, show the potential utility of these methods. In particular, the on-line combination of CITP with tandem MS is shown to offer special promise for sequence determination.
2 Materials and methods 2.1 Samples Biochemical samples were purchased from Sigma (St. Louis, MO, USA) and were used without further purification. Tryptic digests utilized an enzyme-to-substrate ratio of 1:12 in a 1 % ammonium bicarbonate solution at 37 OC for 12 h. Synthetic peptides were kindly provided by Dr. Ian Jardine (Finnigan MAT).
2.2 Instrumentation The CITP experiments utilized simple instrumentation constructed at our laboratory, similar to that described previously for C Z E [ 15-171. These experiments utilized 0.6-2 m lengths of 50 or 100 pm i.d. fused silica capillary tubing suspended between two buffer reservoirs (containing the leading and terminating electrolytes). Both untreated and coated fused silica capillaries were investigated; all results reported were coated capillaries with abondedandcrosslinkedfilmofDB-17 which eliminated electroosmotic flow. The columns were typically rinsed with 1 mL of methanol and conditioned for 10 min by application of 20 kV with the capillary filled with the leading electrolyte. The leading electrolyte in most studies was 10 mM ammonium acetate or potassium acetate/S mM ammonium acetate mixtures (pH 4.5-4.8) prepared in doubly distilled deionized water. The trailing electrolyte was generally 10 m M acetic acid. Reagents were degassed daily by sonication and filtered (0.2 pm) before use. Separations involved two steps. First the sample was introduced as a band (10- I00 nL) at the high voltage end (terminating buffer reservoir) of the capillary and voltage was generally manually ramped over -3 min to the finalvoltagenoted. The high voltage was applied
-
-
-
*
Present address: DOW Chemical, P.O. Box 1398, Pittsburgh, CA 94565, USA 0 173-0835/90/0909 0709 $3.50+.25/0
7 10
R. D. Smith et al.
Electrophoresis 1990.11, 109-711
for typically 10-25 min. The second step involved raising the high voltage reservoir (while continuing to apply high voltages) to cause elution of the seDarated bands Dast the Kratos UV-visible detector (detectioi at 210,254 or i S 0 nm). The detector window was made by removing a small section of the polyimide capillary coating 15-20 cm from the capillary terminus. The window slit width was -2 mm.
2.3 CITP-MS For CITP-MS both 50 pm i.d. untreated and 100 pm i.d. DB- 17 coated (bonded and cross-1inked)fused silica capillary columns were used. Useful results were obtained only with coated columns, as described previously. For most experiments the inlet ofthe capillary was biased to the desiredpotential by an electrode immersed in a buffer solution. The column was loaded with the leading electrolyte, and the head of the capillary and the high voltage electrode (the anode for cationic separations) were placed in the sample reservoir. The sample was loaded into the capillary by electromigration or, preferably, by hydrostatic methods (typically by raising the sample reservoir 4 cm for 30 s). When the desired amount of sample had been loaded, the high voltage was interrupted; the electrode and capillary inlet were then transferred to the trailing electrolyte reservoir, and the voltage reapplied at various rates or steps to begin the separation. The instrumentation for CITP-MS developed at our laboratory has been described elsewhere in detail [ 15-17]. The interface design employs a flowing liquid sheath interface which allow the composition and flow rate of the electrosprayed liquid to be controlled independent of the C E buffer [171. The CITP-MS sheath flow rate (10 % water in methanol at 3 pL/min) and cathode voltage were adjusted to form a stable electrospray at the capillary terminus. The electrical contact is also established through the conductive liquid sheath (typically methanol, acetonitrile, acetone or isopropanol, although small fractions of water or acetic acid are generally added for protein analyses 1301). With this arrangement no significant additional mixing (“dead”) volume (< 10 nL) is produced and analyte contact with metal surfaces is avoided. One problem apparent in the present results is a relatively large variation in ESI signal which apparently resulted from bubble formation near the sheath electrode. The elimination of this problem is the subject of continuing effort. For direct ESI-MS experiments without separation, syringe pumps control the flow of analyte solution and liquid sheath at -0.5 pL/min and 3 pL/min, respectively.
-
2.4 ESI source
The ESI source consists of a 50 or 100 pm i.d. fused silica capillary (the CITP capillary) that protrudes 0.2-0.4 mm from a cylindrical stainless steel electrode[ 171. High voltage, generally -1-4 to +6 kV for positive ions or -5 kV for negative ions, is applied to this electrode. The ESI source (capillary) tip is mounted approximately 1.5 cm from the ion sampling nozzle of the ion sampling orifice (nozzle) of the quadrupole mass spectrometer. A 3-6 L/min countercurrent flow of nitrogen gas is introduced between the nozzle and source to aid desolvation of the highly charged electrospray droplets and to minimize any solvent cluster formation during expansion into the vacuum chamber. Ions enter through the 1 mm diameter orifice and are focused efficiently into a 2 nim diameter skimmer orifice directly in front of the radio frequency (refocusing ouadrupole lens. Typically, + 3 50 V to + 1000Vis
applied to the focusinglens and +200V to thenozzle, while the skimmer is at + 65 V [23,3 11.
3 Results and discussion 3.1 Combination of CE and ESI-MS
Our developmental efforts have, from the outset, focused on the combination of C E and ESI-MS. The history of analytical advances in MS has highlighted the special importance of the combination with separation methods having high selectivity and resolving power in conjunction with the high sensitivity and specificity of MS detection. “Real world” samples are invariable mixtures, often very complex mixtures. Any useful analytical method must accommodate contributions from the sample matrix, interfering substances, etc. The dynamic combination of CE, a separation method of high efficiency and flexibility, with ESI-MS is thus particularly advantageous. Electrophoresis forms a family of related techniques including polyacrylarnide gel electrophoresis, isotachophoresis, isoelectric focusing, gel electrofocusing and two-dimensional electrophoresis for very complex mixtures (i.e. extraordinarly high peak capacity). These methods are of fundamental importance in biochemical analysis where different techniques are applied according to the type of sample and the information required. We have previously described the on-line combination of C Z E with ESI-MS [ 15, 161, the first report of any electrophoretic separation technique in dynamic (i.e. on-line) combination with MS. The development was based upon the recognition that both ends of the C Z E capillary did not have to be immersed in buffer reservoirs, as conventionally practiced. The capillary terminus is adapted as an ESI source which accepts the low (from 0 to 1 pL/min) electroosmotic flow and allows ion production at atmospheric pressure from the electrostatically induced nebulization process.
In earlier versions of the CZE/ESI interface the electrospray dispersion and ionization were accomplished from an electrodeposited metal contact established at the end of the C Z E capillary [ 15, 161. Subsequently, we developed aliquid sheath electrode interface from which the solvent composition and flow rate of the electrospray liquid may be controlled independent of the C Z E buffer (which is desirable since high percentage aqueous and high ionic strength buffers useful in C Z E are not well tolerated by ESI) [ 171. Theelectrical contact is established through the conductive sheath, no additional mixing volumes are produced, no significant positive or negative hydrostatic pressure is produced at the outlet of the electrophoretic capillary (i.e. no induced flow) at normal flow rate of the sheath liquid (1-3 pL/min), and analyte contact with metal surfaces is avoided. This interface provides greatly improved performance and is potentially adaptable to other forms of CE. Because C E relies on analyte charge in solution and the ESI process appears to function most effectively for ionic species, the CE/ESI-MS combination is highly complementary. We have reported the analysis of CZE/MS of a mixture of quaternary ammonium compounds [ 151obtaining over 330 000 plates, an order of magnitude better than obtainable by liquid chromatography (LC) in similar time. In this case sample sizes were 300 femtomoles; however, detection limits of I 10 attomoles are obtainable using single ion detection 1 16I.
-
Capillary isotachophoresis of peptides and proteins
Elecfrophorrsis 1990. 1 1 . 709-111
3.2 Potential of CITP
-w Lys-Trp-Lys
Gly-His-Lys
The steady state concentration ofthe analyte anion for CITP separations is determined by the lead anion concentration and the relative mobilities ofthe solute, lead anion and common cation 1251. Accordingly, if the analyte is much more dilute than the lead anion concentration (which is the case in many analytical separations). the analytes will be concentrated as they separate. In a fully developed separation the concentration of each band is similar and the relative abundance for each component is proportional to the length of the band. Thus, CITP offers the potential for higher sample loading (and increased molar sensitivity), high resolution separations and actual concentration (in many cases) of separated sample bands. Figure 1 shows a typical result of a CITP separation of a simple two-component peptide mixture. In such a case the electrophoretic mobilities of the two cotnponents are quitedifferent and a rapid separation is; obtained yielding near ideal bands. Figures 2 and 3 illustrate separations obtained for an enzymatic digestion of substance P and cytochrome c (primarily due to trypsin but complicated by contamination of chymotrypsin - see later discussion in Section 3.3). In the case of cytochronie c at least 16 peptides are expected, reflected by the complex nature of the isotachopherogram in Fig. 3.
UV 210 nm 0-20 uA in 3 min
i
L_
~
0
min
20
Figure I . CITP separation of two tripeptides.
UV 210 nm ramp to 15 kV
'
11'
40
Figure 2. CITP separation of an enzymatic digest of substance P.
We have previously demonstrated the feasibility of the CITP separation of quaternary phosphonium and ammonium salts, amino acids, and catcholamines 1201 and reference peptides I301 with detection by ESI-MS. Detection limits of approximately 1 0 - I ' have ~ been demonstrated for quaternary phosphonium salts and substantial improvements appear feasible. Sample sizes which can be addressed by CITP are much greater (> 100-fold) than CZE. Samples eluting in CITP are ideally flat-topped bands (the length of the analyte band provides information regarding analyte concentration), well suited to MS as the scan speed (ofthe mass spectrometer need not be challenged by the dynamic nature of "sharp" peaks. Most importantly, however, C l T P provides a relatively pure
MS
I
I
min
0
3.3 CITP with detection by ESI-MS
7 11
GLUCAGON
m/z
34
rnin
44
Figure 3. CITP separation of an enzymatic digest of cytochrome c
MW
Figure 4 . ESI-MS spectrum of glucagon (His Ser Gln Gly 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).
7 12
Electrophoresis lYYO,Il, 709-717
R.D. Smith et a[.
analyte band to the ESI source, without the large concentration of supporting electrolyte demanded by CZE. Thus, CITP/MS has the potential of allowing much greater sensitivities (and analyte ion currents) than may be possible by CZE/MS. For our initial attempt at CJTP-MS of enzymatic digests we selected glucagon, a polypeptide of molecular weight 3483. The features of ESI-MS and ESI-MS/MS for polypeptides are illustrated for the intact glucagon molecule in Figs. 4 and 5. Glucagon (lO-%/L) was introduced as a small band by infusion through the CITP capillary. In Fig. 4 the ESI mass spectrum is dominated by multiply protonated molecular ions of glucagon (3 + to 5 + charge states). A small contribution of a molecular dimer, (2M + 7H)7+,was alsoobserved at m/z 996.
MS/MS
Tryptic Digest Fragments of Glucagon
-
T3
HSQGTFTSDYSK~YLDSR;R !AQDFVQWLMNT TI
TZ
T4
Other enzymatic fragments present (chymotryptic)
-
478'
RAQDFVQW~LMNT 1049
-'
Figure 6 . Expected enzymatic cleavage sites for glucagon.
GLUCAGON
M W 3483
b
B274
m/z Figure 5 . ESI-MS/MS spectra of multiply protonated (3+ to 5 + ) charge states of glucagon (see Fig. 4).
Capillary isotachophoresis of peptides and proteins
Electrophoresis 1990,11, 709-7 17
Figure 5 gives the collision-induced dissociation (CID) mass spectra for each of the 3 + to 5 + charge states of glucagon. The major CID fragments are ascribed to the primary ion
sequence and several assignments are noted in Fig. 5 using the generally accepted nomenclature (augmented by superscript for the fragment ion charge state) 130-321. InspectionofFig. 5
Table 1 . \lalor f‘ragmcnt ions produccd from thc multiply cliarpcd parcnt ions o l glucagon t w p t i c iragments _.____~
_.
Amino acid scq ucnce
Pare t i t ion\ yrodricing :
I’aront prod tieing.
.
I
2 2
[An]
1
[Hll]
2 3 4 5
3
3 3 3
3 3 3 -
.~
I2 Scr GI11
II 10
GIy ‘Ilir
9
6 7 8 9 I0
I’hC Tlir Scr A\P TI 1
7 6 4 3
II
Ser
-7
1 yr
5 4 3
1
2 2
(An] (BnI
I 2 3 4
1.c u
A\P Sc r
2 2 7
3
2 2 2 2 2
3 3 3 3 3 3 3
8
5
3
1 I
2
1
3 3 3 3 ~
2 2 2
2 2 I
2 2 2 2 2
-___
2 2 2 2 2
‘4 -
2 2 >
~.
3
2 2 2 2 2 2 2 2
3 3 3 3 3 3 3
2 2 7 2 2 2 2
I 3 3 3 3 3 3 3
?
3 4 5 6 7 8 9 10 I1
A121 Gln
Asp PilC
Val Gill Try Lcu vet
A\n
11 10 9 8
~
.
~
7 6 5 4 3 2
7 13
2 2 2 7
&
2
3 3
.
_
_
_
7 14
R. 11. Smith e/ ul.
shows that (consistent with other results I3 I I) while some of the amino acid sequence can be deduced (primarily near the C or N termini), large portions of the molecule appear inaccessible using the present experimental C I D conditions. An enzymatic (tryptic) digestion of glucagon would be ex pected to produce four tryptic fragments, as shown by the cleavage site predicted in Fig. 6. Figure 7 shows the total ionization current and UV absorbance (210 and 280 nm) isotachopherograms for the CITP/ESI-MS analysis of a mixture of peptides derived from the digestion with trypsin of glucagon. The time scale shown on the axis refers to the begin ning of the separation. Separation development in CITP may be accomplished before detection starts and, as shown here, the actual time for CITP/MS analysis is determined from the time of application of a hydrostatic head to the trailing electrolyte reservoir to cause elution of the focused bands (here at 27 min). The traces show typical bands and step changes at boundaries between bands. The centers of sample bands show no obvious structure. Strong absorbance at 280 nm indicates the presence of tryptophan in the final band eluting. The total reconstructed ion current for MS detection follows the general features of the UV traces, with better characterized bands apparent at the beginning and the end of the separation and alargely structureless region in the middle. Figure 8 further gives selected ion isotachopherogramsfor the tryptic products arginine(Arg+,m/z 175), T1” (m/z 676) and T2’ (m/z 653) for this analysis. Also detected are additional peptide fragments which arise from the action of contaminating chymotrypsin on the predicted tryptic fragments (T4-478)*’ and (T4-1049)’ and on intact glucagon ( m / z 546 and 366). The mass spectrum for the T1 tryptic fragment of glucagon is given in Fig. 9. The ESI mass spectrum of the TI peptide contains intense 3 + and 2 + peaks, while the singly charged molecular ion is observed to have much lower intensity.
3.4 Primary structural information An important aim of our CZE-MS and CITP-MS efforts is to develop methods that will yield primary structural information (i.e. sequence) for polypeptides and small proteins. The ESI-MS method affords unique opportunities in this regard since ionization efficiencies are high and good results can be obtained even for large proteins 123,311. The fact that ESI mass spectra generally consist of only intact multiply charged molecular ions is sometimes cited as a disadvantage of this method since it is claimed that structural information cannot be obtained. However, as we have shown recently, effective dissociation of molecular ions can be induced in the nozzleskimmer region of the ESI interface [ 3 1,33 I. A more powerful approach is tandem MS of molecular ions from several of the major charge states [31,331. The relatively long and stable period of elution of separated bands in CITP should facilitate MS/MS experiments requiring longer integration, signal averaging or more concentrated samples than provided by CZE. While our current data system prevents obtaining MS/MS spectra of more than one charge state from a single separation, future modifications should facilitate this and more flexible methods for selecting multiple peaks for MS/MS study.
klectrophoresis 1990. 11, 109-7 l i
ENZYMATIC DIGEST OF GLUCAGON 11%;
TOTHL
XZI?
l-i
”1 t
I--
Figure 7. Electrospray total ion current isotachopherogram and corresponding UV absorbance(210and 280nm)tracesfortheCITP/MS analysis of a tryptic digest of glucagon. 2 m x 100 prn column, leading buffer 0.0 1 M ammonium pH 4.9; sample concentration 0.0001 M, 25 YO column volume loading; trailing buffer 0.01 M acetic acid. pH 3.3.
-
The C I D tandem mass spectra for the various tryptic fragments (currently obtained in separate experiments due to limitations of our data acquisition system) can be assembled as previously for the smaller polypeptide, melittin 1321. MS/ MS spectraforthevariouschargestates(1 t t o 3 +)fortheTl fragment ofglucagon are shown in Fig. 10. I n Table 1 we give the charge states of various C1D fragments from MS/MS spectra observed for each of the synthetic tryptic fragments. The C I D fragments observed, primarily due to “B” and “Y” mode cleavage of the polypeptides chain [ 31-331, provide a large amount of redundant information regarding the glucagon sequence. As illustrated by Table 1 , major fragment ions are observed which provide information on the entire glucagon sequence.
4 Concluding remarks C E in combination with MS is making increasing contributions to the study of biopolymers. The MS measurement, the determination of mass via the measurement of the ratio of mass to charge, is qualitatively important but (except for very small species) insufficient for full characterization, given a range of possible charge states. Qualitative investigations of complex systems proceed in combined studies with other techniques, eg., nuclear magnetic resonance. crystallography, etc. Of these, MS is frequently the initial experiment by virtue of its relatively high sensitivity. Similari~. Improved separation methods based upon C E will play an important role in such applications. The concurrent developments in both MS and C E are providing complementary tools for manipula
7 15
Capillary isotachophoresis of peptides and proteins
Electrophoresis 1990, 1 I , 709-7 17
-
#,
mlz
175
mlz
546
I ,i----,i# ,A,# #
1 (74;
Figure 8. Selected electrospray ion current isotachopherograms for the analysis of an enzymatic digest with trypsin (and contaminating chymotrypsin) of glucagon.
30
27
MS
- ,
,
,
,
35
,
,
-mi:, -366
0, ,: ,
40
GLUCAGON T R Y P T I C FRAGMENT T 1
