BIOLOGICAL MASS SPECTROMETRY, VOL. 20, 130-138 (1991)
Mass Spectrometric Quantification of Endogenous P-Endorphin Chhabil Darnap2,Jozef J. Kusmierz' and Dominic M. D e s i d e r i ~ ' , ~ , ~ 'Charles B. Stout Neuroscience Mass Spectrometry Laboratory, and the Departments of *Neurology and 3Biochemistry (D.M.D.),University of Tennessee-Memphis, 800 Madison Avenue, Memphis, Tennessee 38 163, USA
Fast atom bombardment (FAB) mass spectrometry and multiple reaction monitoring (MRM) in the B/E linkedfield scan mode were used to quantify endogenous /&endorphin (BE) in individual human pituitary extracts. The experimental protocol includes the addition of a stable isotope-labeled internal standard ((ZH,-Ile'z)BE,-, to the tissue homogenate before extraction, purification of the native BE by a combination of SepPak chromatography and gradient high-performance liquid chromatography (HPLC), trypin digestion to cleave BE into smaller peptides, and separation of the tryptic fragment BE,,-,, (NAIIK) by isocratic reversed-phase HPLC. Mass spectrometric quantification is based upon recording either (a) the [ M + HI + ions of NAIIK and its deuterated analog (('H4)NAIIK), or (b) the transitions { [NAIIK + HI+ -+ [NAI]'} and {[('H,)NAIIK + HI+ + [('H,)NAI]+} using the B/E linked-field scan. Linear calibration curves were obtained using these two mass spectrometric techniques from standard solutions containing 1.25-20 pg of BE; each standard solution also contained 10 pg of ('H,)BE. The amounts (% f sd.)of endogenous BE in five separate human pituitaries were found to be 156 84 ([ M HI method) and 169 & 99 pmol mg- protein (MRM method).
+
+
'
phins and to elucidate the roles that they play in pathophysiologies, it is critical to determine their amounts in various body tissues. Highly structure-specific and sensitive analytical methods are needed to identify and B-Endorphin (BE) is an endogenous opioid peptide that quantify endogenous peptides so that the data reflect consists of 31 amino acid residues (see Scheme 1) and with high fidelity the amino acid sequence of the target has a molecular weight (mol. wt.) of 3463. BE belongs to peptide. one of the three known major groups of endogenous The present study was undertaken with the objective opioid peptides, the other two being the enkephalins of developing a quantification method capable of proand dynorphins. BE is biosynthesized endogenously in viding a sensitive and sequence-specific measurement of neurons by proteolytic cleavage of the large gene endogenous BE in extracts of human tissues. Fast atom product precursor pro-opiomelanocortin (POMC), bombardment (FAB) mass spectrometry is the method which is also the precursor of several other biologically of choice for the analysis of highly polar and fragile active peptides such as /%lipotropin (B-LPH), corticocompounds such as pep tide^.^ This gentle mode of iontropin (ACTH), and several melanotropins (MSH).' BE ization has the unique potential to generate sequencecomprises the C-terminal segment of POMC, and is the specific data for the peptide being measured by only component of the precursor peptide with agonist providing the [M + H I + ion and amino acid sequenceactivity at an opioid receptor (epsilon). Further metadetermining fragment Tandem mass spectrombolic processing of BE produces its C-terminally shortetry (MS/MS),' which provides structure-specific data ened homologs such as a-endorphin, y-endorphin, by establishing a unique precursor-product relationBE,-,, ,and their N,-acetylated forms.' ship, can be used to quantify a target peptide by meaP-Endorphin has been implicated in several neurochemical and pharmacological functions of the b ~ d y . ~ , ~suring the ion current due to a selected amino acid sequence-determining fragment ion formed from the To understand the physiological significance of endormass-selected molecular ion (i.e., the [M + H I + ion) of that peptide.".' experimental strategies are discussed here : in 1Tyr-Gly-Gly-Phe-Met-Thr-Ser-Glu-9Lys1 one,Twoendogenous BE in a human pituitary extract homogenized with the internal standard is purified by a combination of Sep-Pak chromatography and gradient reversed-phase high-performance liquid chromatography (RP-HPLC). Trypsinolysis produces a smaller peptide BE',-'4 (i.e., the pentapeptide having the sequence NAIIK). NAIIK is separated from the other tryptic peptides by isocratic HPLC, and the FAB mass Scheme 1 . Amino acid sequence of 8-endorphin. The highspectrometric ion current due to the [M H I + ions of The arrows indilighted segment is the tryptic fragment BE,. cate the five trypsin cleavage points. that tryptic fragment and its deuterated analog INTRODUCTION
I
+
1052-930619 11030130-09 $05.00
0 1991 by John Wiley & Sons, Ltd.
Received 21 September 1990 Accepted (revised) 26 November 1990
QUANTIFICATION O F ENDOGENOUS /%ENDORPHIN
(NA('H,-I)IK) are compared. In the second technique, the same experimental protocol is followed, with the exception that mass spectrometric measurements are done by recording in a multiple reaction monitoring (MRM) mode the area under the ion current profiles of the reactions { [NAIIK] + H] + [NAI] +} and {[('H,)NAIIK + H I + -,[(2H,)NAI]+} that take place in the first field-free region (FFR) of a forward-geometry double-focusing mass spectrometer. The utility of these two separate analytical methods is demonstrated by analyzing the endogenous BE content in several individual human pituitary samples. Most laboratories use radioimmunoassay (RIA) to analyze endogenous BE."-14 Although RIA possesses detection sensitivity at the femtomole level, its putative molecular specificity, which is generally attributed to binding of an antibody that is directed towards a specific amino acid sequence within an antigen, is questionable' ' because that specific sequence may also be present in other homologs, precursors, and metabolites of that antigen. The RIA data sometimes are reported as 'like immunoreactive', which cannot correlate unambiguously to the sequence of the target peptide. For example, the antisera used by Rossier et al.', and Vuolteenaho et a l l 4 are stated to be 'specific' , and yet those antisera cross-reacted towards BE,,,, fully with several related peptides such as 8-LPH, a precursor protein, BE,-27, N,-acetylated BE and BE,-27 because they also contain that common internal peptide sequence. +
EXPERIMENTAL Materials Synthetic BE, glutathione, dithiothreitol (DTT), dithioerythritol (DTE), and 3-nitrobenzyl alcohol (NBA) were purchased from Sigma Chemical Co. (St Louis, Missouri), and were used without further purification. The purity of BE was verified by RP-HPLC before use. ('H,)Ile (Merck, St Louis, Missouri), di-tertbutyl dicarbonate (Fluka Chemical Co., Ronkonkoma, New York) and mercaptoacetic acid (Aldrich Chemical Corp., Milwaukee, Wisconsin) were purchased. Solid-phase peptide synthesis (SPPS) of the deuterated internal standard, ((2H,)Ile22)BEl~,l,human, was performed using Boc chemistry on an Applied Biosystems (Foster City, California) Model 430 solid-phase peptide synthesizer. Boc-('H,)Ile was prepared according to the literature procedure (yield 98%).15 All other amino acids and reagents were purchased from Applied Biosystems. The H F treatment (performed by Immunodynamics, San Diego, California) of the final SPPS human sulfoxide product provided (ZH,-IleZZ)BEl~,l, ((2H,-IleZZ)BE[0],-,l, human), which was reduced to (ZH,-Ile22)BEl-31~ with mercaptoacetic acid using the procedure of Houghten and Li.16
Tissue extraction and purification Human pituitaries were obtained at autopsy, and were stored at - 70°C until processed."
131
Whole pituitaries were homogenized individually for 30 s in cold (4 "C) 1 M acetic acid (1 : 20, w/v) with a Polytron (Brinkman, Westbury, New York) at setting 6 (13OOO r.p.m.) as described previou~ly.'~ The deuterated internal standard was also added at the same time. After equilibration of the exogenous peptide with endogenous peptide for 2 h at 4 "C, the homogenate was centrifuged for 30 min at 131OOO x g. The supernatant was subjected to a ODS RP-Sep-Pak cartridge (Waters Associates, Milford, Massachusetts)," which was activated previously by washing successively with 4 ml of methanol, 4 ml of water, and 8 ml of 0.1% trifluoroacetic acid (TFA). Salts and water-soluble organics were removed by washing the cartridge with 4 ml of 0.1% TFA, and the peptide-rich fraction was eluted with 4 ml of a CH3CN:0.1% TFA solvent mixture (equal volumes), and lyophilized to dryness. The lyophilizate was dissolved in 150 pl of 0.1 % TFA and was resolved into individual fractions by a gradient ODS RP-HPLC system, which consisted of a Varian Model 5000 microprocessor-driven high-performance liquid chromatograph outfitted with an ODS pbondapak (10.4 x 300 mm, 10 pm bead diameter, 30 A pore size; Waters Associates) analytical steel column and an ultraviolet detector (210 nm). One-minute fractions were collected across a 90 min gradient that combined several linear segments, and a volatile triethylamine formate (TEAF; 40 mM; pH 3.1) buffer and CH,CN organic m~difier.'~ The retention time of the column was calibrated with synthetic BE, which eluted at 73 min. Before analyzing the pituitary extract, the column was washed extensively with CH,CN (loo%, 20 min) to ensure complete removal of synthetic BE. The BE fractions from the pituitary extracts were collected at 73 f 1 min, and lyophilized to dryness. Figure 1 depicts a representative reversed-phase HPLC trace of an extract of one of the post-mortem pituitaries (B).
Enzymolysis The solutions of the peptide standards used for obtaining the calibration curves (see below) and the gradient HPLC BE lyophilizate of the pituitary extracts were treated separately with trypsin. Samples were dissolved in 4 ml of 50 mM tris(hydroxymethy1)aminomethane (Tris; pH 7.6) and incubated for 30 min at 20°C with immobilized trypsin beads (0.14 units; Pierce, Rockford, Illinois), which were washed previously (four times) with Tris and centrifuged before use. Trypsinolysis was stopped by removing the beads by centrifugation. To purify the tryptic fragment NAIIK, the supernatant was eluted from an ODS Sep-Pak cartridge and from a reversed-phase HPLC column (6% CH,CN in TEAF). Trysinolysis was nearly quantitative ( > 99%)''
Mass spectrometry A VG 70E-HF forward-geometry double-focusing mass spectrometer was used to obtain the FAB mass spectral data. The instrument was outfitted with a VG FAB ion source, an IonTech BIINF saddle-field FAB gun, and a
C. DASS, J. J. KUSMIERZ A N D D. M. DESIDERIO
132
BE
0.64
v)
L
3
a
0
J
,
I
20
40
60
T i m e (min)
Figure 1. A gradient RP-HPLC trace of an extract of pituitary B. For further analysis, the fraction eluting within *1 min of the peak indicated by the arrow was collected.
DEC PDP-11/24 minicomputer-based 11-250 M+ data system. Xenon atoms of approximately 7 keV impact energy and 1 mA emission current served as the ionizing beam. The FAB matrix consisted of a 3: 1 (w/w) mixture of DTT and DTE. The isocratically purified NAIIK was dissolved in 100 pl of 0.1% TFA. An aliquot (generally 1 11) of that solution was deposited onto the FAB probe tip, which also contained 0.4 p1 of the FAB matrix. Addition of glutathione (1 nmol) to the probe tip enhanced the [M + HI' ion current of NAIJK." The FAB-produced ions exited the ion source at a potential of 6 kV. The [M + HI' ion data for NAIIK and ('H,)NAIIK were obtained by scanning the magnet across the mass range of 554-561 u at a scan rate of 5 s per decade, and by operating the data system as a multichannel analyzer (MCA); several scans (50) were accumulated. The MRM measurements of the first FFR reactions of {[NAIIK + HI+ + [NAI]'} and the corresponding deuterated analog were made with a 250 ms ion collection time and a 20 ms intermass delay time. Although the instrument was adjusted to a resolution of 500 (10% valley definition), the effective resolution of the precursor ion mass-selection is lower (100-200) using a linkedfield scan at constant B/E (where B and E are magnetic and electric fields, respectively). Collision-activated dissociation (CAD) was accomplished using helium as the neutral target; the pressure of helium was adjusted to reduce the precursor ion's intensity to 70% of its original value. The ion current due to the CAD reactions was monitored for 1 min, and the area under those ion current profiles was used to construct the calibration curve. The B/E linked-field spectra of the [M + H]+ ion of the tryptic fragment NAIIK obtained from synthetic BE and from one of the post-mortem pituitaries (B) were acquired with an exponential-down scan of 5 s per decade in the mass range of 10-650 u. Seven scans were signal-averaged using the MCA to obtain the final spectra shown below. Helium was used as the CAD gas. To obtain calibration curves, two sets of standard solutions were prepared. One set contained 20, 10, 5, 2.5, and 1.25 pg of BE and a fixed amount (10 bg) of ('H,)BE. The second set contained 2000 to 62.5 ng
(2000, 1000, 500, 250, 125 and 62.5 ng) of synthetic BE and a constant amount (1 pg) of the internal standard. The entire content of these solutions was trypsinized separately as described above, and NAIIK was purified by isocratic HPLC. The lyophilized fractions were dissolved in 100 p1 of 0.1%TFA. Three FAB mass spectrometric measurements were made by separate 1 p1 loadings of each of the standard solutions. ~
~~~~
RESULTS AND DISCUSSION
A more complete knowledge at the molecular level of the role of neuropeptides in various biochemical processes demands their quantification in body tissues and fluids at femtomole levels with unambiguous sequencespecific analytical methods. For example, the anatomic and metabolic relationships among the three opioid peptidergic pathways (namely, endarphinergic, enkephalinergic and dynorphinergic), where each system is comprised of a large precursor, several intermediatesized precursors, bioactive peptides, and their inactive metabolites, further stress the importance of unambiguous sequence-specific analytical data. The analysis is complicated further by the fact that several enzymes operate within those peptidergic pathways to produce structurally closely related peptides. Therefore, the primary objective of this research was to develop and to use a highly structure-specific mass spectrometric method for quantification of endogenous BE in human tissue extracts. That type of analytical data can be obtained by use of MS/MS, which embraces the concept that MS-1 mass-selects the precursor ion of interest and the sequence-specific fragment ions of that precursor are recorded by MS-2. A B/E linked-field scang also embodies a similar concept, and links incontrovertibly the product ions to the precursor ion, within the limits of effective precursor ion resolution. The use of a suitable internal standard is another prerequisite for the success of any quantification method because it compensates for any inconsistency of the various experimental steps of the analytical procedure.
133
QUANTIFICATION OF ENDOGENOUS 8-ENDORPHIN
The internal standard must possess chemical and physical properties identical to those of the native peptide to minimize any differences in tissue extraction, chromatographic separation, and mass spectrometric response. A stable isotope-labeled analog of the native peptide ideally meets these criteria. We have synthesized human by replacing Ile (residue 22) in (ZH4-Ile23)BEl-31, BE with ~,~-2,3,4,4-(’H,)Ileto shift the mass of the tryptic fragment NAIIK by 4 u. Chemical analogs of a native peptide have‘ been used as an internal standard.21-24However, because the peak intensities of the [M H]+ ion in an FAB mass spectrum are a function of the relative hydrophobicities of the various constituents present on the FAB probe tip, the differences in an FAB desorption profile result in inconsistent peak ratios when a chemical analog of a peptide is used as an internal standard.” A stable isotope-labeled internal standard overcomes that pr~blem.’~ An analytical measurement using an intact oligopeptide rather than one of its smaller fragments certainly possesses a higher level of molecular specificity (e.g. BE versus NAIIK). However, several experimental difficulties are encountered in the analysis of intact BE by FAB mass spectrometry. First, the mol. wt. of BE is beyond the most sensitive mass range of many commercial mass spectrometers. Second, bulky molecules such as BE have a reduced probability of desorption by bombardment of an energetic particle beam, and also have a reduced detection efficiency by a conventional electron multiplier detector. However, ion formation and ion detection can be improved to a certain extent by the use of a cesium ion beamz6 and a multichannel array detector,27 respectively. Third, the relative hydrophilic nature of BE, as measured by its Bull and Breese indexz8(B & B index = +930 for BE),” also limits efficient FAB ionization of intact BE. Fourth, a higher mass resolution (approximately 4000) is required to resolve the individual ions in the molecular ion cluster of BE. Because transmission of ions in a mass spectrometer is inversely proportional to the mass resolution, a corresponding reduction in detection sensitivity results when the mass-resolved [M + HI’ ion of intact BE is used for quantification. Finally, MS/MS becomes ineffective to quantify large molecules like BE because of poor fragment ion yields, which are caused by dissipation of energy into the large number of internal degrees of freedom available during the collision activation pro~ess.’~ Another approach to analyze compounds of high mass is to consider multiply charged ions, which extends the effective mass range of a mass spectrometer. For example, a quantification method utilizing the [M 5HI5+ ion formed by electrospray of BE was reported recently by us.30 Although that method exhibits an excellent detection sensitivity (5 fmol) for synthetic BE, its current limitation is the nonavailability of a suitable stable isotope-labeled internal standard. For example, incorporation of more than ten deuterium atoms is required to mass-separate the [M 5HI5+ions of the native peptide and the isotopelabeled internal standard. Voyksner and PackZ1 analyzed BE in human cerebrospinal fluid (CSF) extracts by monitoring the [M 2H]’+ ion formed with thermospray mass spectrometry and by using its
+
+
+
+
chemical analog, N,-acetyl-BE, as an internal standard to achieve a detection sensitivity of 10 pmol. However, no specific sequence ions were reported. FAB of BE also produces doubly charged ions. However, our attempts to use the doubly charged molecular ions for FAB mass spectrometric quantification of BE were not encouraging. Recording the [M 2H]’+ ion at 1732.5 u across a narrow 10 u mass window with NBA containing 1 p1 of 0.1% TFA and 2.5 pg glutathione was found to be sensitive down to 3 pmol of synthetic BE. However, even when half of the amount of an extract from five combined human pituitaries was analyzed, no unambiguous evidence of endogenous BE was noted. Although high ion current was obtained at the mlz of the doubly charged ion of BE from that biological extract, the BE signal could not be differentiated from the background because several constituents still present in that extract (an observation based upon the electrospray mass spectrometric analysis of another pituitary extract)30 were ionized with ease. Such an experience indicates the importance of using MS/MS (the B/E mode here) to minimize the background chemical noise and, consequently, to improve the signal-to-noiseratio. Alternatively, enzymolysis of BE into smaller segments for FAB mass spectrometric analysis is an attractive proposition. An experimental protocol that uses trypsinolysis of BE to produce six smaller peptides separates these tryptic peptides into the individual components, and employs FAB mass spectrometry of those components, was developed recently by us to characterize endogenous BE.” The tryptic pentapeptide NAIIK was found to exhibit a significantly enhanced FAB mass spectrometric response (approximately 1340 times versus intact BE). The hydrophobic character (B & B index = -188),” the smaller size of NAIIK and the additional isocratic HPLC purification step all combined to contribute to that enhanced FAB mass spectrometric response. Because no stable isotopeincorporated internal standard was available at that time, quantification of endogenous BE was not attempted. In the present study, that strategy has been extended further to develop a quantification method for endogenous BE in biological samples by using (2H4-IleZZ)BEl-,l,human as the peptide internal standard. An additional improvement includes the use of the B/E linked-field MRM to increase the molecular specificity of the measurements. Figure 2 illustrates the overall experimental protocol employed.
+
Quantification by monitoring the [ M NAIIK
+ H J + ion of
The [M + H] ion is generally the most abundant ion in the FAB mass spectrum of a pe~tide.~-’A similar observation was also made in the FAB B-scan spectrum (not shown) of NAIIK. Therefore, monitoring the [M HI+ ion must be considered first for sensitive quantitative measurements. The set of standard solutions that contained 10 pg of the internal standard was analyzed for the [M + H]+ ions of NAIIK and (’H,)NAIIK. The plot (not shown) of the ratio (an average of three measurements) in arbitrary units of ion current due to those [M + H]+ ions +
+
C. DASS, J. J. KUSMIERZ AND D. M. DESIDERIO
134
extraction/purification steps was 20%," and that the entire sample was loaded on the FAB probe tip.
Pituitary Tissue
t
Homogenization (add cold acetic acid, ['H&BE)
Quantification by multiple reaction monitoring
t Sep-Pak ODS Chromatography (elution with 1:l CH3CN:0.1% CF3COOH)
+
t
Gradient ODS RP-HPLC (elution with TEAF, 40 mM, pH 3.1 and CH3CN)
$.
Trypsin Digestion
I
Isocratic ODS RP-HPLC (elution with 6% CH3CN)
t t FAB Mass Spectrometry
Collect NAIIK and ['H,]-NAIIK
J
Monitor the [M + HI' ions of NAIIK and [*H,]-NAIIK
\ Monitor the transitions ([NAIIK + HI' {(12H41-NAIIK
+
-
NAIf}and H)' [*H 41 N A I '}
-
Figure 2. The overall experimental protocol used to quantify native BE.
against the amount of BE was linear (correlation coefficient r2 = 0.999). The standard deviation of the three measurements was within k 1%. That calibration curve was used to quantify endogenous BE in human pituitary extracts. The method is sensitive down to 86 fmol of synthetic BE (as NAIIK) present on the FAB probe tip." The corresponding detection limit for a biological sample is estimated to be approximately 500 fmol, assuming that the recovery of the native BE from the
-
Although monitoring the [M + H] ion of NAIIK provides a means to quantify BE with high detection sensitivity sufficient for many physiological measurements, it nonetheless lacks the molecular specificity needed for accurate measurement of endogenous neuropeptides because an impurity in a biological sample may also produce a signal at the m/z of that [M + H]+ ion, and because the [M + H I + indicates the molecular weight of a peptide but not its amino acid sequence. As stated above, an accurate measurement can be made only by monitoring a sequence-specific fragment ion from a specifically selected precursor ion. We have used the BIE linked-field scan to monitor the reaction {[NAIIK + H I + + [NAI]+}. The B / E linked-field scan maintains throughout the experiment a rigorous structural correlation between the precursor ion and its product ions; that specific precursor-product link increases significantly the specificity and the accuracy of the resulting quantitative measurements. The Nterminal B, fragment [NAI]' (m/z 299) is the most abundant fragment ion in the B / E CAD spectrum of NAIIK (see Fig. 3), and therefore is the logical choice for quantitative measurement of endogenous BE (as NAIIK). The standard nomenclature of peptide bond fragmentation is used in Fig. 3 to designate the amino acid sequence-specific fragment ions.31 The standard solutions of BE that contained 10 pg of (2H,-Ile22)BE,-3,, after trypsinolysis and isocratic HPLC separation of NAIIK, were analyzed by X65.0-
X65.0
8
100-
(M+H)+
A
80 Lys-OH
444
373
260
147
Y"
60.
*i" 5
I
2!9
1 400 500 Figure 3. The 6/€linked-field scan of the [M + H]+ ion at m/z 558 of synthetic NAIIK.
1 -
135
QUANTIFICATION OF ENDOGENOUS 8-ENDORPHIN
mass-selecting the precursor ions at m/z 558 and 562 (the [M + H I + ions of NAIIK and (2H4)NAIIK, respectively) in the first FFR, and recording the ion current due to their product ions [NAI]' and [(2H4)NAI] at m/z 299 and 303, respectively, formed by CAD of those precursors. The calibration curve (not shown) obtained by plotting the ratio of the areas under the ion current profiles observed in these CAD fragmentation reactions versus the corresponding amount of synthetic BE is linear (correlation coeffcient I' = 0.999). Each data point of the curve is an average of three measurements, and has a precision of < f 3%. In a separate experiment, the detection sensitivity of this analytical method was demonstrated to be 500 fmol of synthetic BE (as NAIIK) present on the FAB probe tip. +
Analysis of endogenous /?-endorphin in human pituitaries 8-Endorphin is one of the products of the metabolic processing of POMC, which is synthesized primarily in the cell body of neurons located in the pituitary gland.' Metabolic defects in a peptidergic pathway may be correlated with various clinical and pathological manifestations. In order to understand the functional significance of BE in various pathophysiologies, it is important to establish the BE levels in the pituitary of normal and diseased ( e g tumorous pituitaries) human subjects. Towards that goal, a study was initiated to measure the human pituitary levels of BE using the mass spectrometric techniques discussed above. Before quantifying endogenous BE (as NAIIK) in these pituitary extracts, it was important first to prove qualitatively that NAIIK was present. Thus, a complete
B/E linked-field scan of one of the pituitary extracts (B) was obtained (see Fig. 4). The close correlation of the spectra in Fig. 4 (biological) and Fig. 3 (synthetic), within the typical fO.5 u in these B/E linked-field scans," proves that this isocratic tryptic fragment is NAIIK. To achieve greater accuracy in the analytical measurements, the amount of the internal standard should be comparable (generally within the range of one-fourth to four times) with the endogenous level of the native peptide. In a preliminary study, 1 pg (290 fmol) of the deuterated internal standard was used. Endogenous BE was quantified in each pituitary extract by using the two mass spectrometric methods described above. The amount of BE was calculated by using calibration curves obtained by analyzing the set of standard solutions that contained 1 pg of ('H4-Ile2')BE,-,,, human. The three human pituitaries contained 9.6, 16.4 and 16.0 pg of BE (equivalent to 93, 69 and 65 pmol BE mg-' protein), respectively, by the [M + H] ion monitoring method. The corresponding amounts of BE found by the MRM method were 13.9, 19.6 and 18.1 pg (equivalent to 135, 82 and 73 pmol BE mg-' protein), respectively. Those two sets of data do not differ significantly. However, it is obvious from the above data that the amount of internal standard added was less than 10% of the amount of endogenous BE, and was thus outside the ideal range stipulated above for accurate quantitative measurements. Therefore, for analysis of a next set of five human pituitaries, 10 pg of (ZH4-Ile22)BEl-,,,human was added to each pituitary. To demonstrate typical signal-to-noise levels observed, the +
x22.0 .9
x22.0
-
100,
8.
I
I
0
100
200
a00
329
I 1 1
I I II.
I
.
ll
Ill 500
Figure 4. The B/€ linked-field scan of the m/z 558 from a tryptic fragment of an extract of pituitary B. The fraction (indicated by an arrow in Fig. 1 ) eluting at 73 f 1 rnin was trypsinized, and the tryptic fragment NAIIK, after isocratic RP-HPLC purification, was used to obtain this spectrum.
C. DASS, J. J. KUSMIERZ AND D. M. DESIDERIO
136 NAllK A
i\
r2 H, ] - NAllK
+
f Pituitary C
A
Pituitary
D
+
-.-/---------
--lLdL-
Pituitary E
555
560
adds a significantly increased level of molecular specificity to the analysis of endogenous neuropeptides. At one level, the [M HI+ ion data indicate that a peptide with an appropriate mol. wt. is being quantified. At a higher level of specificity, the MRM data indicate that the appropriate mol. wt. also has an appropriate subsequence attached to it. Finally, a complete B / E linked-field scan confirms unambiguously the identity of NAIIK, as is demonstrated in Fig. 4. For many biological samples (such as an extract of CSF), the detection sensitivity is an important factor; however, sensitivity is not a problem with pituitaries where nanomole amounts of BE are available. However, the molecular specificity is of utmost importance for all biological samples because of their multi-component nature, and more importantly because the sequence of the target peptide must be known. It is also possible that an HPLC emuent, which contains HPLC column bleed and biological matrix that elute with NAIIK, might also produce a signal at the m/z of the [M H I + ion of NAIIK. In addition, the use of an FAB liquid matrix results in peaks at every mass unit, which could also contribute to the NAIIK [M + H I + ion signal. These interferences will be very critical when dealing with low amounts of endogenous material, and may cast suspicion on the validity of the [M + H]+ ion measurements. It must also be remembered that the MCA mode of data acquisition will improve signal-toelectronic noise but not the signal-to-chemical noise. This latter type of noise can be minimized readily by the use of the MS/MS technique (here, MRM). Because of all of these considerations, we prefer the MRM method for quantification of endogenous neuropeptides. In FAB mass spectrometry, the ionization is controlled largely by the surface activity of the solute. Hydrophobic peptides are more surface-active and thus are ionized preferentially at the expense of hydrophilic peptides. Therefore, to achieve the optimal detection sensitivity, it is important that the sample being quantified be highly homogeneous. The Sep-Pak chromatography used here removes salts and water-soluble polar organic compounds. The gradient reversed-phase HPLC resolves the Sep-Pak peptide-rich fraction into individual fractions. Because the resolution of RP-HPLC is less than infinite and because a biological extract is multi-component in nature, it is likely that some other substances present in the pituitary extract might co-elute with BE.30 Therefore, isocratic RP-HPLC of the trypsin digest is an essential step to provide a cleaner sample for FAB mass spectrometric analysis. This isocratic step not only removes the interferences of those co-eluting substances, but it also isolates NAIIK from other tryptic fragments, which otherwise would interfere in FAB ionization of NAIIK. The pentapeptide NAIIK is comprised of four different amino acids. These amino acids can rearrange differently to provide 4! = 24 peptides. That number becomes staggering when other amino acids that have the same nominal mass as one of these four amino acids (e.g., Ile = Leu and Asn = Orn; Orn, however, is not a naturally occurring amino acid) are involved in these permutations. Because of differences in their hydrophobicity, those other peptides would exhibit different chromatographic behavior, and may not pass through the
565
Figure 5. Scan of the [M + H I + region (554-567 u) of NAllK and (2H,)NAIIK from extracts of five human pituitaries A-E.
[M + H I + ion MCA data for the five pituitary extracts A-E are presented in Fig. 5. The amount of BE was calculated by using the calibration curves obtained by analyzing the set of standard solutions that contained 10 pg of the internal standard. The results, each an average of duplicate measurements, are included in Table 1. The average amounts of BE found in five human pituitaries by the two mass spectrometric techniques ([M + H I + ion and MRM) were 32 f 23 pg and 35 f 26 pg, respectively, equivalent to 156 f 84 pmol mg-' protein (or 17 f 9 nmol BE g-' tissue) and 169 f 99 pmol mg-' protein (18 f 10 nmol BE g-' tissue), respectively (2 f s.d.). The coefficient of variation in each set of data is a reflection of variation of the peptide in the biological tissue, and not of the analytical methods, which, as reported above, have precisions of f 1% and f3%, respectively. The data also demonstrate that the addition of 10-15 pg of deuterated internal standard is appropriate for quantification of endogenous BE in human pituitaries. Vuolteenaho et a1.14 found whole human pituitaries to contain 84 f 17 nmol g-' tissue (n = 10) 'BE-like' immunoreactivity (BE-LI), of which 1 5 4 5 % was found to be 'BE-like' material. In another RIA study, 101 human pituitaries were found to contain 18.2 nmol g-' tissue BE-LI.32 The RIA of a whole rat pituitary gland extract has been shown to contain 77.7 f 5.8 nmol g-' tissue BE-LI.I3 Our results obtained by the two mass spectrometric techniques are compatible with those measurements. Although RIA provides a very high level of detection sensitivity, and in combination with HPLC reduces the possibility of cross-reactivities with potential interfering substances, it cannot provide the amino acid sequence of the target peptide." The use of mass spectrometry
137
QUANTIFICATION O F ENDOGENOUS &ENDORPHIN
Table 1. Amount of native findorphin in post-mortem human pituitaries Amount
[M + H I + measurements Pituitary
A
6 C D
E Average
Weight (a)
Protein
(me)
0.605 0.595 0.640 0.280 0.470
68.9 75.2 62.0 30.2 48.1
MRM measurements
w3
nmol g-' tissue
pmol mg-' protein
R
nmol g-' tissue
pmol mg-' protein
17.1 54.8 58.6 13.9 15.1
8.2 26.6 26.4 14.4 9.3
72 21 1 273 133 91
17.2 57.5 67.6 14.3 17.0
8.2 27.9 30.5 14.7 10.4
72 221 31 5 136 102
32*23
17*9
156*84
35*26
18*10
169*99
The [ M + H]+ and M R M measurements are precise within *1% and *3% (n = 3), respectively.
final stage of analysis. For example, a pentapeptide NALLK (B & B index = -268),, elutes 2.2 min after NAIIK (B & B index = -188) on our isocratic reversed-phase high-performance liquid chromatograph. It is for these reasons that our analysis scheme has used a multi-facet approach: gradient RP-HPLC to ensure that the fraction collected is endogenous BE, trypsinolysis of that fraction to ensure that NAIIK is produced, isocratic RP-HPLC to ensure that the fraction eluting is NAIIK, a full-scan B/E linked-field spectrum to confirm the sequence of that tryptic fragment, and the MRM measurement to ensure a link between WAIIK + H I f and [NAI]'. All of these steps must be performed in each analysis, and they contribute to the fact that the results indeed are a true representation of the endogenous amount of BE. We have selected a B, N-terminus ion in MRM measurements, and thus 3! = 6 subsequences are possible. However, to date, this present system has the highest level of molecular specificity to quantify endogenous BE.
The multi-facet approach employed here provides a significantly increased level of molecular specificity of the method. Instability of ion currents inherent with the FAB mode of ionization demands the incorporation of an internal standard in the analysis scheme. The use of a deuterated analog as an internal standard compensates for any experimental inconsistency. The principles underlying this study are being used in other studies. For example in one study, endogenous BE and methionine enkephalin (ME) were analyzed simultaneously in post-mortem control pituitaries and in pituitary tumors. The preliminary results, which will be reported elsewhere, suggest that the difference in the BE content between controls and tumors is significant. That finding indicates a clear-cut aberration of POMC processing in pituitary tumors. In a second study, four separate quantification methods were compared for endogenous ME.34 Tissues were divided into four portions, and radioreceptorassay (RRA), RIA, mass spectrometry ([M H]+ monitoring) and MRM data were accumulated. The different quantitative data demonstrated a ratio of 8 : 2 : 3 : 3 for RRA:RIA: [M H]+ :MRM methods, respectively.
+
+
CONCLUSIONS This study has demonstrated that FAB mass spectrometry in combination with the B/E linked-field MRM is a suitable analytical method to quantify endogenous BE.
Acknowledgements This research was supported by the NIH grants (DMD) GM 26666 and RR 01651.
REFERENCES 1. S. Numa, in Peptides, Vol. 6, ed. by S. Udenfriend, p. 1. Aca-
2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
demic Press, New York (1984). D. G. Smyth, D. E. Massey, S. Zakarian and M. D. A. Finnie, Nature 279,252 (1979). T. L. O'Donohue and D. M . Dorsa, Peptides 3, 353 (1982). H. Khachaturian, M . E. Lewis, T. Kang and S. J. Watson, in Handbook of Chemical Neuroanatomy, Vol. 4, ed. by A Bjorklund and T.Hokfeldt, p. 21 6. Elsevier, Amsterdam (1985). M . Barber, R. S. Bordoli, R. D. Sedgewick and A. N. Tyler, J. Chem. SOC.Chem. Commun. 325 (1981). K. Biemann and S. Martin, Mass Spectrom. Rew. 6, 1 (1987). C. Dass and D. M . Desiderio,Anal. Biochem. 163, 52 (1987). C. Dass, in Mass Spectrometry of Peptides, ed. by D. M. Desiderio, p. 327. CRC Press, Boca Raton (1991). F. W. McLafferty, Tandem Mass Spectrometry. Wiley, New York (1983). D. M . Desiderio and C. Dass, Anal. Lett. 19,1963 (1986). D. M . Desiderio, Analysis of Neuropeptides by Liquid Chromatography and Mass Spectrometry. Elsevier, Amsterdam
12. 13. 14. 15. 16. 17.
18. 19. 20.
(1984); D. M . Desiderio (ed.),Mass Spectrometry of Peptides. CRC Press, Boca Raton (1991). F. Nyberg, A. Wahlstrom, 6.Sjolund and L. Terenius, Brain Res. 259, 267 (1987). J. Rossier, T. M. Vargo, S. Minick, N. Ling, F. E. Bloom and R. Guillemin, Proc. NatlAcad. Sci. USA 74, 5162 (1977). 0. Vuolteenaho, J. Leppaluoto, M . Hoyhtya and J. Hirvonen, Acta Endocrinol. 102, 27 (1983). M. Itoh, D. Hagiwara and T. Kamiya, Tetrahedron Lett. 4393 (1975). R. A. Houghten and C. H. Li,Anal. Biochem. 98,36 (1979). J. J. Kusmierz, R. Sumrada and D. M . Desiderio, Anal. Chem. 62,2395 (1 990). T. Higa and D. M. Desiderio, lnt. J. Pept. Prot. Res. 33. 250 (1989). G. H. Fridland and D. M . Desiderio, J. Chromatogr. Biomed. Appl. 379,251 (1986). C. Dass, G. H. Fridland, P. W. Tinsley, J. T. Killmar and D. M. Desiderio, lnt. J . Pept. Prot. Res. 34, 81 (1989).
138
C. DASS, J. J. KUSMIERZ AND D. M. DESIDERIO
21. R. D. Voyksner and T. W. Pack, Biomed. Environ. Mass Spectrom. 18,897 (1989). 22. E. Benfenati, G. Icardi, S. Chen and R. Fanelli, J. Labeled Compd. Radiopharm. 28,411 (1990). 23. D. M. Desiderio, S. Yarnada, F. S. Tanzer, J. Horton and J. Trirnble, J. Chromatogr. 217,437 (1981). 24. S. Yamada and D. M. Desiderio, Anal. Biochem. 127, 213 ( 1 982). 25. D. M. Desiderio and M. Kai, Biomed. Mass Specfrom. 10, 471 (1 983). 26. W. Aberth, Anal. Chem. 58,1221 (1986). 27. J. S. Cottrell and S. Evans, Anal. Chem. 59,1990 (1987).
28. M. B. Bull and K Breese, Arch. Biochem. Biophys. 161, 665 (1975). 29. P. J. Todd and F. W. McLafferty, Int. J. Mass Spectrom. Ion Phys. 38,371 (1981). 30. C. Dass, J. J. Kusrnierz, D. M. Desiderio, S. A. Jarvis and B. N. Green, J. Am. SOC.Mass Specfrom. (1991). 31. P. Roepstorff and J. Fohlrnan, Biomed. Mass Specfrom. 11, 601 (1984). 32. 0. Vuolteenaho, Acta Physiol. Scand. Suppl. 531, 1, (1 984). 33. B & B index was calculated according to ref. 28. 34. J. Lovelace, J. J. Kusmierz and D. M. Desiderio, J. Chrumarogr. Biomed. Appl. 530 (235252) 1991.