De novo sequencing by nano-electrospray multiple-stage tandem ...

N. Carte et al., Eur.J. Mass Spectrom. 7, 399–408 (2001)

399

De Novo Sequencing of an Immune-Induced Peptide of Drosophila Melanogaster N. Carte et al., Eur.J. Mass Spectrom. 7, 399–408 (2001)

De novo sequencing by nano-electrospray multiple-stage tandem mass spectrometry of an immune-induced peptide of Drosophila melanogaster

Nathalie Carte, Nukhet Cavusoglu, Emmanuelle Leize* and Alain Van Dorsselaer Laboratoire de Spectrométrie de Masse Bio-Organique, UMR 7509, ECPM, 25 rue Becquerel, 67087 Strasbourg Cedex, France. E-mail: [email protected]

Maurice Charlet and Philippe Bulet* Institut de Biologie Moléculaire et Cellulaire, UPR 9022, CNRS, “Réponse Immunitaire et Développement chez les Insectes”, 15 rue René Descartes, 67084 Strasbourg Cedex, France. E-mail: [email protected]

To combat infection, the fruit fly Drosophila melanogaster responds by rapid synthesis of a series of immune-induced molecules reported as Drosophila immune-induced molecules (DIMs). Characterization of the primary structure of the DIMs is required to establish their exact function. In order to get such information, previous studies on the elucidation of primary structures of the DIMs were developed using a methodology combining matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS), highperformance liquid chromatography (HPLC), enzymatic digestion and Edman degradation. Nevertheless, some of the DIMs were resistant to classical Edman sequencing. Therefore, mass spectrometry was used to characterize the primary structure of one of the DIMs, namely the N-blocked DIM13 peptide. The complete sequence of DIM13 was established by means of a strategy of nanoelectrospray ionisation (ESI) combined with multiple-stage tandem mass spectrometry (MSn) and then was partially confirmed with a combination of enzymatic digestions and MALDI-MS analyses. Interestingly, most of the amino acid sequences have been deter3 4 2 mined using three-stage (MS ) and four-stage (MS ) tandem experiments, whereas classical tandem mass spectrometry (MS ) yielded only incomplete sequence information. Finally, DIM13 is a 23 amino acid peptide with a pyroglutamic modification at the N terminal position. This work illustrates the remarkable advantages of MS3 and MS4 compared with the MS2 experiment for de novo peptide sequencing. The use of nano-ESI also makes these experiments compatible with the low amount (picomolar level) of DIM13 peptide available for sequencing by ESI-MSn. n

Keywords: nano-electrospray, ion trap MS , de novo sequencing, Drosophila melanogaster, innate immunity

Introduction The fruit fly Drosophila melanogaster, like other insects, is particularly resistant to microbial infections. Three main innate immune mechanisms are considered to contribute to this resistance: (1) cellular reactions allowing phagocytosis and encapsulation of the microorganisms by the insect blood cells (hemocytes), (2) activation of proteolytic cascades leading to melanization, coagulation and opsonization, and (3) transient activation of the synthesis of antimicrobial peptides.1 Recent advances in the molecular mechanisms of innate host defence reactions in

multicellular organisms from mammals to insects, including Drosophila, have outlined important similarities. For this reason and because of the potency of Drosophila for genetic studies, this model appears as a powerful tool to study the evolution of the mechanisms of innate immunity. Although the structure and the regulated expression of the antimicrobial peptides have been intensively studied during 2,3 the past ten years, only fragmentary data are available on the other immune processes developed by the fruit fly to fight off an infection. In order to find new Drosophila immune molecules, not related to antimicrobial substances, a protocol consisting of a differential display by matrix-

© IM Publications 2001, ISSN 1356-1049

400

De Novo Sequencing of an Immune-Induced Peptide of Drosophila Melanogaster

assisted laser desorption/ionization mass spectrometry (MALDI-MS) and high-performance liquid chromatography (HPLC) has been optimized on Drosophila hemolymph (insect blood) collected from experimentally infected and non-infected flies.4 MALDI-MS analysis led to the detection of at least 24 immune-induced molecules (DIMs, i.e. Drosophila immune-induced molecules) in the Drosophila 4,5 hemolymph, 24 hours following an experimental infection. Some of the DIMs were resistant to classical Edman sequencing degradation, for example DIM13, which was found to be N-terminal blocked. Moreover, the chromatographic fraction of DIM13 was not pure enough for use of the classical approach, which comprises enzymatic cleavage of the peptide followed by tandem mass spectrometry (MS/MS) experiments of the resulting lower mass peptides. A better purity could be achieved, but to the detriment of the amount of DIM13 available for the sequence determination. In our study, we intended to demonstrate that a strategy relying on the multiple-stage tandem mass spectrometry (MSn) 6 capabilities of an ion trap mass spectrometer could be used to determine de novo sequence of the entire DIM13 peptide. For the structural elucidation of the DIM13 peptide, threestage (MS3) and four-stage (MS4) experiments performed with an ion trap mass spectrometer yielded significant and relevant sequence information to fully characterize this Nblocked peptide. As we were sequencing a novel peptide, further analyses, such as reduction, alkylation and enzymatic digestion using a carboxypeptidase followed by MALDI-MS were performed to partially validate the sequence of DIM13 peptide determined by electrospray ionzisation (ESI)-MSn. In n order to perform all the ESI-MS experiments with the estimated 10 pmoles of DIM13 available, we took advantage of the low consumption of analyte using the coupling of the nano-electrospray source with an ion trap mass analyzer.7–9 The elucidation of the primary structure of this 23 amino acid peptide DIM13 is presented here in detail.

Material and methods Insect immunization and hemolymph collection

The insect immunization and blood collection were per4 formed according to the procedure previously described. HPLC separation and Edman degradation

The acidified hemolymph collected from 180 flies was subjected to separation using reversed-phase (RP)-HPLC on an Aquapore RP 300 C8 column (1 × 100 mm). Separation was performed with a linear gradient of 0–80% acetonitrile in acidified water [0.05% trifluoroaceticacid (TFA)] over 80 min at a flow rate of 80 µL min–1 at 35°C. The fractions, which were collected by hand in low-protein-binding Eppendorf tubes, were analyzed by MALDI-MS in order to detect the fraction containing DIM13. A second purification step was performed with a linear gradient from 2% to 22%

acetonitrile in acidified water over 10 min and from 22% to 37% in 75 min at a flow rate of 80 µL min–1 on the same column as previously at 35°C. The HPLC fraction, which did not contain the DIM13 peptide exclusively, was estimated, according to its optical density, at 40 pmoles. This estimation was done by comparison with the optical density of peptides of the same family. We divided the total amount in three aliquots: one for the Edman degradation (10 pmoles), n one for the nano-ES-MS experiments (10 pmoles) and the last one for enzymatic digestions and MALDI-MS analyses (20 pmoles). An Edman degradation experiment was attempted on the HPLC fraction containing mainly DIM13. Automated Edman degradation of DIM13 was performed on a pulse liquid automatic sequenator (Applied Biosystems ABI473A). Nano-electrospray multiple-stage tandem mass spectrometry (NanoES-MSn) Sample preparation

To avoid the inconvenience of the TFA present in the HPLC sample during mass spectrometric analyses, the dissolved DIM13 (10 pmoles estimated) was loaded on a C18 Zip Tip (Millipore) and concentrated in 2 µL H2O + CH3CN (50 : 50 with 1% formic acid). All tandem mass spectrometric analyses were performed with a total volume of 2 µL of a concentrated DIM13 solution, deposited on a previously rinsed nanospray capillary. Electrospray instrumentation

Low-energy collisions of multiply-charged ions were performed on an ion trap mass spectrometer (ESQUIRE-LC, Bruker–Franzen Analytik GmbH, Germany) equipped with a nanospray source. The gold/palladium-coated nanospray capillaries were from Protana (Odense, Denmark). The capillary and the counter-electrode voltages were set to 750 V and 250 V, respectively. The voltage applied on the exit of the metallized-glass capillary interface was optimized at 80 V. The voltage applied on skimmer 1 was optimized at 40 V. Calibration of the mass analyzer was performed with the multiply-charged ions of the following five standard peptides: leu-enkephalin, angiotensin, substance P, bombesin, and ACTH, having monoisotopic molecular weights of 711.38, 1045.54, 1346.74, 1619.81 and 2464.20 Da, respectively. To perform tandem mass spectrometry experiments n (MS ), isolation of the precursor ion was achieved by scanning frequencies of ions to eject all other ions from the trap. The precursor ion was fragmented by applying a resonance frequency on the end cap electrodes (peak-to-peak amplitude of 0.8 to 2.5 V) matching the frequency of the selected ion. As a result, the kinetic energy of the precursor ion increases and dissociation, due to collisions with the helium buffer gas (pressure of 5 × 10–3 mbar) occurs. Sequences of isolation and fragmentation were repeated for MSn experiments to gain structural information on the selected frag-


n+1

ment ions. In this study, to obtain MS (n = 1, 2, 3) fragmentation mass spectra of a good quality, the precursor n ion was chosen according to its intensity on the different MS n–1 or MS spectra. Ions were scanned in standard resolution mode with a scan speed of 13,000 m/z per second. A total of 20 scans were averaged to obtain a mass spectrum.

Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS)

Sample reduction and alkylation

About 20 pmoles of DIM13 was dissolved in 25 mM NH4HCO3 buffer at pH 8. First, the peptide was reduced in the presence of 10 mM of Dithiothreitol (DTT, Sigma, St Louis, USA) in a 200-molar excess at 57°C for one hour. Second, alkylation was performed with a 240-molar excess of 55 mM iodoacetamide (Sigma, St Louis, USA) as alkylating agent.

401

MALDI instrumentation

This study was carried out on a BIFLEX matrix-assisted laser desorption/ionization time-of-flight mass spectrometer (Bruker, Bremen, Germany), equipped with SCOUT highresolution optics and a gridless reflectron. This instrument has a maximum acceleration potential of 20 kV and may be operated in either linear or reflectron mode. Ionization is accomplished with the 337 nm beam from a nitrogen laser with a repetition rate of 3 Hz. A camera mounted on a microscope allows visualization of the sample crystallization homogeneity before measurements. The spectra were acquired in the positive-ionization mode using both the linear and reflectron modes. External calibration was performed with five standard peptides: leu-enkephalin, angiotensin, substance P, bombesin and ACTH, with average/monoisotopic molecular weights of 711.82 / 711.38, 1046.2 / 1045.54, 1347.66 / 1346.74, 1620.86 / 1619.81 and 2465.71 / 2464.20 Da, respectively.

Results and discussion Trypsin digestion

An aliquot of the reduced and alkylated DIM13 (10 pmoles) dissolved in 25 mM NH4HCO3 buffer at pH 8 was subjected to trypsin digestion (sequencing-grade trypsin, Promega) using an enzyme-to-substrate ratio of 1 : 20 (w / w). 0.5 µL of this solution was deposited on the MALDI probe after two hours for overnight incubation at 35°C. Carboxypeptidase P digestion

An aliquot of the reduced and alkylated DIM13 (10 pmoles) dissolved in 25 mM ammonium carbonate buffer at pH 8, was treated with carboxypeptidase P (Boehringer, Mannheim, Germany; initially stored in 25 mM sodium citrate) at an enzyme-to-substrate ratio of 1 : 20 (w / w).10–12 0.5 µL of the product was deposited on a MALDI probe (see following section) after 30 s, 2 min, 5 min, 10 min and 48 h of incubation at 37°C. This C-terminal sequencing by carboxypeptidase P was followed by MALDI-MS and gave optimal sequencing data after 48 hours of incubation.

The strategy used to determine the de novo sequence of the entire Drosophila melanogaster immune-induced pepn tide DIM13, relied on the MS capabilities of the ion trap mass spectrometer. Thus, MALDI-MS was used on peptide digests to confirm the determined sequence. Additional information on DIM13 was also obtained after reduction–alkylation experiments, indicating that no cysteine residue was present in the peptide sequence. The nano-ES spectrum of the HPLC fraction containing DIM13 displayed + three major charge states at m/z 1325.7 (2 ) (monoisotopic + + mass), m/z 884.3 (3 ) and m/z 663.8 (4 ) (average values) giving a measured monoisotopic weight value for DIM13 of 2649.4 ± 0.2 Da (Figure 1). In agreement with the mobile-proton model,13 different collision-induced dissociation (CID) processes using ESn MS are known to occur according to the charge state of the

Sample deposition for MALDI analysis

The thin-layer sample preparation method was used for MALDI-MS analysis. Briefly, 0.5 µL of α-cyano-4–1 hydroxycinnamic acid (HCCA, Sigma; 7 g L in acetone) was placed on the probe tip. When this was dry, 0.5 µL of 1% TFA was deposited on the crystallized matrix bed, followed by the deposition of 0.5 µL of digestion mixture loaded at different time intervals. The acidity of the matrix and of the 1% TFA droplet was sufficient to completely quench any further digestion. After drying, the target was washed with 2 µL of 1% aqueous TFA, the liquid removed after a few seconds using forced air and the sample again dried under vacuum.

Figure 1. NanoES-MS spectrum of the DIM13 peptide dissolved in H2O + CH3CN (50 : 50) with 1% of HCOOH.

402


Figure 2. MS2 fragment-ion spectra obtained from the DIM13 peptide : (a) (2+) molecular ion at m/z 1325.7, (b) (3+) molecular ion at m/z 884.3 and (c) (4+) molecular ion at m/z 663.8.

14

2

peptide precursor ion. Consequently, MS spectra were performed successively on each multiply-charged ion observed in the DIM13 nano-ES spectrum to obtain as much structural information as possible. The MS2 spectrum of the (2+) molecular ion (m/z + 1325.7) [Figure 2(a)] displayed 10 significant singly- (1 ) + and doubly- (2 ) charged fragment ions (Table 1). The C-terminal amino acid, a phenylalanine, was recognised by inter+ pretation of the (2 ) fragment ions at m/z 1243.1, m/z 1234.2

and m/z 1225.1, corresponding respectively to b22, b22–H2O and b22–2H2O, and was further confirmed by the fragmenta+ tion of the (3 ) molecular ion (m/z 884.3). The phenylalanine was placed at the C-terminus since the N-terminus of the DIM13 is expected to be blocked on the basis of the failed attempt at Edman degradation. The MS2 spectrum of the (3+) molecular ion (m/z 884.3) [Figure 2(b)] gave 11 significant fragments ions (Table 1) whose interpretation did not allow any further sequence


403

Table 1. Masses and interpretations of DIM13 fragments obtained with MS2, MS3 and MS4 experiments. Fragments by MS

2

Fragments by MS +

Parent ion: 1325.7 (2 ) + 1764.7 (1 ) b15–NH3 +

1658.8 (1 ) y15 +

1463.7 (1 ) b12

+

Parent ion: 992.3 (1 ) b8 +

+

1307.7 (2 ) M+2H –2H2O + 1243.1 (2 ) b22 +

1234.2 (2 ) b22H2O + 1225.1 (2 ) b22–2H2O +

1187.6 (1 ) y11 +

Fragments by MS +

Parent ion: 1187.6 (1 ) y11 +

974.3 (1 ) b8–H2O + 920.3 (1 ) a8–CO2

1170.6 (1 )y11–NH3

+

1022.4 (1 ) y10–H2O + 983.5 (1 ) y9

+

877.2 (1 ) b7 +

966.4 (1 ) y9–NH3

+

869.4 (1 ) y8

780.3 (1 ) b7 763.3 (1 ) b6–NH3 +

734.4 (1 ) HVERPD + 624.4 (1 ) b5

+

352.2 (1 ) a3–16 + 241.7 (1 ) b2–H2O

+

(ϕ–CH=CH–NH–CH2–CH2–

+

imidazole+H )

+

+

812.3 (1 ) y7 +

720.2 (1 ) FGNGGFSA–H2O

+

992.3 (1 ) b8

597.3 (1 ) VERPD + 580.3 (1 ) VERPD–H2O

+

869.5 (1 ) y8

+

495.2 (1 ) b4 +

467.2 (1 ) a4 +

396.2 (1 ) b3 +

368.2 (1 ) a4 +

Parent ion: 884.3 (3 ) + + 1325.6 (2 ) M+2H +

1187.6 (1 ) y11

+

992.3 (1 ) b8

1040.4 (1 ) y10 +

400.3 RTVD–CO2–CO + 357.2(1 ) RTV

+

258.2 (1 ) RT

+

780.3 (1 ) b6

992.3 (1 ) b8 +

+

Parent ion: 472.2 (1 ) RTVD + 428.2 (1 ) RTVD–CO2

+

840.3 (1 ) RPDRTVD + 822.2 (1 ) RPDRTVD–H2O

+

+

948.3 (2 ) b17 +

+

Parent ion: 732.3 (2 ) b12

+

846.7 (3 ) M+3H –Z + 834.3 (2 ) b14 +

829.7 (2 ) y15

723.2 (2 ) b12–H2O + 683.3 (2 ) c11

+

Parent ion: 846.7 (3 ) + * 1237.5 (1 ) b 11 (b11–Z)

+

674.8 (2 ) b11

+

+

1187.6 (1 ) y11

+

1119.4 (2 ) y20–NH3

+

1019.7 (1 ) b 9–H2O(b9–H2O–Z) + 973.2 (1 ) HVERPDRT–H2O

+

934.7 (1 ) VDFGNGGFSA–NH3

660.8 (2 ) a11

+

732.3 (2 ) b12

+

624.2 (1 ) b5

+

624.2 (1 ) b5

616.2 (2 ) b10–H2O + 603.2 (2 ) HVERPDRTVD

+

472.2 (1 ) RTVD

+

*

+

495.2 (1 ) b4 +

472.2 (1 ) RTVD + 467.2 (1 ) a4

DRTVDFGNG–CO +

679 (2 ) b

* 12

(b12–Z)

+

396.1 (1 ) b3 +

Parent ion: 663.8 (4 ) + 1021.9 (2 ) b18 +

948.3 (2 ) b17 +

891.3 (2 ) b15 +

+

1401.7 (1 ) ERPDRTVDFGNGG + 1383.4 (1 ) ERPDRTVDFGNGG–H2O +

+

884.5 (3 ) M+3H

+

Parent ion: 948.3 (2 ) b17

1272.6 (1 ) RPDRTVDFGNGG + 1254.6 (1 ) VERPDRTVDFG–H2O

919.8 (2 ) b16 +

+

+

1197.6 (1 ) VERPDRTVDF–H2O +

1116.3 (1 ) PDRTVDFGNGG + 1099.6 (1 ) PDRTVDFGNGG–NH3

842.7 (2 ) c14 +

834.2 (2 ) b14 +

+

829.4 (3 ) b22

1098.4 (1 ) PDRTVDFGNGG–H2O + 1019.4 (1 ) DRTVDFGNGG

+

777.1 (3 ) b21 +

+

1002.4 (1 ) PDRTVDFGN + 992.3 (1 ) b8

755.4 (1 ) y6 +

752.2 (3 ) y20 +

+

734.7 (3 ) b20

974.3 (1 ) b8–H2O

+

PDRTVDFGN–28

711.2 (3 ) b19 +

+

682.1 (3 ) b18 +

+

659.4 (4 ) M + H –H2O + 632.8 (3 ) b17 +

613.8 (3 ) b16

934.7 (1 ) DRTVDFGNG –28 + 904.3 (1 ) RTVDFGNGG +

894.4 (1 ) c7 +

887.3 (1 ) RTVDFGNGG–NH3

+

608.3 (1 ) y5 +

594.8 (3 ) b15 +

521.3 (1 ) y4 +

450.3 (1 ) y3

DRTVDFGN–H2O +

780.3 (1 ) b6 +

732.3 (2 ) b12

+

Parent ion: 396.2 (1 ) b3 379.9 (1 ) b3–17 + 368.0 (1 ) a3

+

+

4

+

1040.5 (1 ) y10

894.4 (1 ) c7

+

3

+

723.3 (2 ) b12–H2O + 713.8 (2 ) VERPDRTVDFGNGG–NH3 +

701.3 (2 ) ERPDRTVDFGNGG + 683.3 (2 ) c11 +

674.8 (2 ) b11 +

665.8 (2 ) b11–H2O + 624.2 (1 ) b5 +

495.2 (1 ) b4 +

467.2 (1 ) a4 +

396.1 (1 ) b3

404


Scheme 1. DIM13 sequence established (a) from MS2 spectrum of the (4+) molecular ion at m/z 663.8 and (b) from MS2 spectrum of the (3+) molecular ion at m/z 884.3. Note that bold annotations represent the main cleavages.

determination. Nevertheless, a triply-charged fragment ion at m/z 846.7, corresponding to the loss of 112.8 Da from the molecular mass of the DIM13 peptide, was submitted to collision-induced dissociation since the results revealed a pyroglutamic acid located on the N-terminal end of the DIM13 peptide (vide infra for MS3 on the m/z 846.7 fragment confirming the N-termination). 2 Finally, the MS spectrum obtained for the fragmenta+ tion of the 4 molecular ion (m/z 663.8) was more complex [Figure 2(c)] and displayed singly-, doubly- and triplycharged fragment ions (Table 1). Sets of ions displaying the same charge state were, however, able to provide internal sequence information. Thus, the F18 S19 A20 tag was identified + + by interpretation of the (1 ) fragment ions. From the (2 ) fragment ions, the internal sequence N15 G16 G17 F18 could be + deduced. Finally, the (3 ) fragment-ion series allowed elucidation of the sequence at the C-terminal end. The C-terminal sequence of DIM13 was found to be in agreement with the former internal sequence tags determined from MS2 spectra. 2 According to the whole MS results, the C-terminal sequence of the DIM13 was identified as (b14= 1667.4) – N15 G16 G17 F18 S19 A20 Q/K21 R22 F23 [Scheme 1(a)]. The DIM13 sequence orientation (N- or C-terminus) 2 deduced by MS was confirmed by carboxypeptidase P treatment and analysis by MALDI-MS. To follow the C-terminus enzymatic sequencing, the kinetics of digestion were established (see Experimental). Figure 3 shows the MALDI mass spectrum from the sequential degradation of DIM13. Carboxypeptidase treatment on the singly-charged (1+) molecular ion of DIM13 at m/z 2650.0 (average value) gave a first peak at m/z 2502.9 corresponding to a loss of 147.1 Da

confirming the presence of a phenylalanine as the C-terminal amino acid. The successive mass losses of 156.0, 127.9, 70.8 and 87.0 Da confirmed the C-terminal sequence for DIM13 as S19 A20 Q/K21 R22 F23. To identify the complete sequence of DIM13, the determined C-terminal sequence (b14=1667.4)–N15 G16 G17 F18 S19 A20 Q/K21 R22 F23 was searched in protein databases (Proteinprospector http://prospector.ucsf.edu). No successful identification, at the date of the search (September 1999),

Figure 3. MALDI-MS mass spectrum of the carboxypeptidase digestion of DIM13 after 48 hours of incubation. Note that the observed sodium adducts were due to the sodium citrate buffer in which the carboxypeptidase P was initially stored.


405

Figure 4. MS3 fragment ion spectra obtained from the DIM13 peptide : (a) (1+) precursor ion at m/z 992.3, (b) (2+) precursor ion at m/z 948.3 and (c) (1+) precursor ion at m/z 1187.6.

was found. To characterize the full primary sequence of this 3 4 novel peptide, MS and MS tandem mass spectrometry was required to generate sequence information from the uncharacterized (N-terminal) region of the DIM13 peptide. The MS3 experiments performed on the (1+) precursor ion at m/z 992.3 essentially generated N-terminal fragment ions [Figure 4(a) and Table1] from which a partial sequence on the N-terminal side of DIM13 was obtained, viz. (b3=396.2)–V4 E5 R6 P7 D8 [Scheme 2(a)]. The loss of CO2

from the a8 ion corroborates the presence of acidic amino acids (D and/or E) in the fragment at m/z 992.3. In the same manner, ion b6 loses NH3 and, therefore, indicates a basic amino acid such as the arginine located in the fragment at + m/z 992.3. Moreover, low intensity (1 ) ions at m/z 734.4, m/z 597.3, m/z 580.3 were identified as internal fragments of the precursor at m/z 992.3 and indicated the presence of a histidine residue on the N-terminal side of DIM13 (H3 V4 E5 R6 P7 D8).

406


Scheme 2. DIM13 sequence established from MS3 spectra of precursors: (a) (1+) fragment ion at m/z 992.3, (b) (2+) fragment ion at m/z 948.3.

3

+

In addition, MS analysis performed on the (2 ) ion at m/z 948.3 [Figure 4(b) and Table 1] led to three major frag+ + + ments at m/z 992.3 (1 ), m/z 904.3 (1 ) and m/z 732.3 (2 ). It is important to notice for the further attribution of minor ions + that the fragmentation of the (2 ) precursor ion at m/z 948.3 + yields the two (1 ) fragment ions at m/z 992.3 and m/z 904.3 by the same peptide bond cleavage [Scheme 2(b)]. Thus, minor ions, previously identified as the daughter ions of m/z 992.3, attend their complementary fragment ions from the m/z 948.3 doubly-charged precursor ion [Scheme 2(b)]. This attribution unequivocally confirmed the sequence on the Nterminal side. MS3 analysis performed on the (2+) precursor ion at m/z 732.3 exhibited one main cleavage giving two complemen-

tary (1+) fragment ions at m/z 992.3 and m/z 472.2. In addition, the series of b ions, formerly characterized as daughter ions of m/z 992.3 [Scheme 3(a)], was also observed. Minor fragments were further interpreted thanks to the elucidation of the previously uncharacterized (1+) fragment ion at m/z 472.2. 4 For sensitivity reasons, MS could not be performed on this fragment ion at m/z 472.2. So, the precursor ion at m/z 2 + 472.2 was directly isolated from the MS of the (3 ) molecu+ lar ion at m/z 884.3, which also displayed a (2 ) fragment ion + at m/z 732.3 and a (1 ) fragment ion m/z 992.3. Therefore, the isolated ion at m/z 472.2 was presumed to be obtained by a double fragmentation resulting from the successive cleavages of m/z 884.3 and m/z 732.3. The MS3 experiment per-

Scheme 3. DIM13 sequence established from MS3 spectra of precursors: (a) (2+) fragment ion at m/z 732.3, (b) (1+) fragment ion at m/z 472.2 and (c) (1+) fragment ion at m/z 1187.6.


formed on the fragment ion at m/z 472.2 displayed low intensity fragments that gave the internal sequence R9 T10 V11 D12 [Scheme 3(b)]. Attribution of the arginine as the last amino acid of the fragment m/z 472.2 was deduced by a calculation based on the mass difference between the precursor ion and the identified sequence tag T10 V11 D12 [472.2 – 101(T) – 99(V) – 115(D) = 157.2 which corresponds to + (R + H) ]. Tryptic digestion of DIM13 and analysis by MALDI-MS confirmed this attribution. Characterization of the fragment at m/z 472.2 allowed 3 the interpretation of minor ions obtained from the MS of the + (2 ) precursor ion at m/z 732.3 which were attributed as internal fragments containing the sequence R9 T10 V11 D12 [Table 1 + and Scheme 3(a)]. Moreover, the (2 ) fragment ion at m/z 603.2 identified as H3 V4 E5 R6 P7 D8 R9 T10 V11 D12, confirmed the presence of H3 on the N-terminal side of the DIM13 peptide. The last N-terminal amino acids containing the N-terminal blocking group were deduced from the MS2 and MS3 results as follows. The N-terminal blocking group was assumed to be a pyroglutamic acid (Z in the one-letter code) as this is frequently observed for small bioactive peptides and supported by the loss of 112.8 Da from the m/z 846.7 ion. To confirm this result, an MS3 experiment was performed on the ion at m/z 846.7. The obtained mass spectrum displayed ions (Table 1) formerly characterized as y-ions and internal-fragment ions. Moreover, b-fragment ions (denoted as b* fragments) corresponding to original b-fragment ions minus the mass of a pyroglutamic residue (112 Da) were also observed. Therefore, since no b-ions of the DIM13 peptide were observed, the precursor ion at m/z 846.7 was unequivocally assigned to be DIM13 without its pyroglutamic N-termination. Considering the pyroglutamic acid N-termination and the smallest b-ion (b3 at m/z 396.2) detected in the fragmentation spectra, we could identify the amino acid located at the second position of the N-terminal side. Thus, the b3-ion at m/z 396.2 was shown to contain a histidine residue (137 Da) plus the pyroglutamic acid N-termination (112 Da), so that a phenylalanine (147 Da) was deduced as the missing amino acid on the N-terminal side of DIM13 peptide. To confirm the presence of a phenylalanine at the second position, the b3-ion at 396.2 m/z was further fragmented (MS4). The corre4 sponding MS spectrum exhibited one major fragment at m/z 241.7, which justified the attribution of the phenylalanine. The C-terminal sequence of DIM13 was elucidated 3 + using the MS fragmentation from the (1 ) precursor ion at 3 m/z 1187.6 [Figure 4(c) and Scheme 3(c)]. The MS spectra displayed series of y ions (Table 1) and identified the C terminus as F13 G14 N15 G16 G17 F18 S19 A20 Q/K21 R22 F23 [Scheme 2 3 4 3(c)]. All the interpretations from the MS , MS and MS analyses led to the following DIM13 sequence: Z1 F2 H3 V4 E5 R6 P7 D8 R9 T10 V11 D12 F13 G14 N15 G16 G17 F18 S19 A20 Q/K21 R22 F23 In our sequence determination, we still had an ambiguity for the attribution of the residue at position 21, which

407

could be either a glutamine or a lysine amino acid. These isobaric residues can be differentiated under high-energy collision experiments, thanks to the formation of a specific 15 16 d-fragment ion. More recently, Bahr et al. have described n a methodology, based on MS experiments performed using an ion trap, which allows the differentiation of glutamine from lysine. By sequential MSn experiments on b- or y-fragment ions adjacent to the ambiguous residue, a peptide tag containing a glutamine residue loses a formamide molecule (45 Da) from the lateral chain, which is specific enough to differentiate the two isobaric residues. Accordingly, we attempted an MS3 experiment on the y3-ion at m/z 450.3, which is a fragment ion adjacent to the ambiguous residue. However, the spectrum obtained did not allow a determination of the residue at position 21 as described by Bahr et al.16 Therefore, to elucidate the nature of the ambiguous residue at position 21, trypsinolysis of the HPLC fraction containing DIM13 with other co-eluted peptides, was performed. At this stage of the work, the sequence of DIM13 was known and it was, therefore, possible to clearly identify the DIM13 tryptic peptides among the others. The enzymatic digest analyzed by MALDI-MS notably showed the following identified ions for the DIM13 peptide: m/z 1166.2 (Z1 F2 H3 V4 E5 R6 P7 D8 R9), m/z 1355.6 (T10 V11 D12 F13 G14 N15 G16 G17 F18 S19 A20 Q/K21 R22) and m/z 1503.7 (T10 V11 D12 F13 G14 N15 G16 G17 F18 S19 A20 Q/K21 R22 F23). A lysine in position 21 would have led to tryptic fragments at m/z 2348.5, m/z 1568.6 and m/z 1200.3. These results allowed us to identify a glutamine residue at position 21, if no missed cleavages occurred. The elucidation of the primary structure of DIM13 was complex since MS2 experiments generated only a few fragn ments and the use of MS (n > 2) was therefore necessary to recover the whole sequence. The fragmentation behavior of the DIM13 peptide is considerably different from that of tryptic peptides which end with an arginine or a lysine and which are known to fragment easily thanks to the permanent charge located on the C-terminus.17 As the DIM13 peptide + + contains 23 amino acids, high charge states (4 and 3 ) are required to have enough energy to fragment this large peptide. Fragmentation of high-charge-state ions generates daughter ions of the same charge state, or lower, which has to be precisely determined to interpret the mass spectra. The ion trap analyzer offers sufficient resolution to distinguish the different charge states and was therefore appropriate to sequence 4+ and 3+ ions for the characterization of the DIM13 peptide. 2 From the interpretation of all the MS mass spectra, 38% 3 of the DIM13 sequence was elucidated. The data from MS 4 and MS spectra, performed using chosen fragment-ion precursors characterized the entire primary structure of this 2 peptide. Nevertheless, the few fragments obtained by MS for this peptide correlate with previous studies showing that the observed fragmentations depend on the nature and 18 charge state of the precursor ion of the peptide. Confirmation using model peptides showed that the number of ioniz-

408


ing protons relative to the number of basic residues present in peptides containing aspartic acid (D) or glutamic acid (E) residues influences the dissociation patterns of protonated peptides. Specific and dominant cleavages at acid residues were shown to occur when the number of ionizing protons equals the number of arginine residues in these peptides. On the contrary, when the number of ionizing protons is greater than the number of arginine residues, non-specific cleavages are observed. According to previous observations, the MS2 experi+ ments of the (3 ) molecular ion (all the ionizing protons are located on the three arginine residues) exclusively displayed series of b and y ions adjacent to the acidic amino acid residues at m/z 624.2 (1+) b5, m/z 992.3 (1+) b8, m/z 732.3 (2+) b12 + + and m/z 829.7 (2 ) y15, m/z 1187.6 (1 ) y11 as major fragments 2 [Scheme 1(b)]. On the other hand, MS fragmentation per+ formed on the (4 ) molecular ion (the ionizing protons are now more numerous than the number of arginine residues) exhibited a more complex tandem mass spectrum with nonspecific fragmentations that was, nevertheless, richer in sequence information [Scheme 1(a)].

Conclusions This work shows that the complete sequence determination of a novel natural 23-amino-acid peptide with a blocked N-terminus may be achieved by a combination of MSn analyses performed on an ion trap analyzer, MALDI-MS with enzymatic digests and the use of a fast HPLC procedure. It is significant to point out that MS3 and MS4 experiments pro2 vided key sequence information whereas MS experiments yielded only a small amount of sequence information. Finally, using this structural information, searches in the Berkeley Drosophila Genome Project data base (BDGP) using the TBLASTN program now found two nucleotide sequences encoding DIM13 which were located on the same chromosome arm (2R) in 50A9–50A9 according to the fix annotations (CG18278, CG18279) from the Genome Annotation Database of Drosophila (GadFly). The sequence of the DIM13 peptide has been patented on 29 November 2000 under the number 00/15434. The total identification of the two nucleotide sequences suggests a gene duplication. In addition, in the vicinity of DIM13, two additional isoforms were observed with 81% and 68% of amino acid sequence similarities.

Lutte contre la Mucoviscidose (AFLM) and NIH (1PO1 AI44220-02). N.C. thanks the Bruker Daltonics Society and the CNRS for financially co-supporting her PhD fellowship.

References 1. M. Meister, C. Hetru and J.A. Hoffmann, in Current Topics in Microbiology and Immunity, Ed by L. Du Pasquier and G.W. Litman. Springler-Verlag, Berlin, Germany, 248, 17 (2000). 2. B. Lemaitre, Médecine/Sciences 15, 15 (1999). 3. P. Bulet, Médecine/Sciences 15, 23 (1999). 4. S. Uttenweiler-Joseph, M. Moniatte, M. Lagueux, A. Van Dorsselaer, J.A. Hoffmann and P. Bulet, Proc. Natl. Acad. Sci. USA 95, 11342 (1998). 5. P. Bulet and S. Uttenweiler-Joseph, in Proteome and Protein Analysis, Ed by R.M. Kamp, D. Kyriakidis and Th. Choli-Papalopoudos. Springer-Verlag, Berlin, Germany, 157 (1999). 6. K.R. Jonscher and J.R. Yates, Anal. Biochem. 244, 1 (1997). 7. M. Wilm and M. Mann, Anal. Chem. 68, 1 (1996). 8. M.R. Emmet and R.M. Caprioli, J. Am. Soc. Mass Spectrom. 5, 605 (1994). 9. D.C. Gale and R.D. Smith, Rapid Commun. Mass Spectrom. 7, 1017 (1993). 10. D.H. Patterson, G.E. Tarr, F.E. Regnier and S.A. Martin, Anal. Chem. 67, 3971 (1995). 11. B. Thiede, J. Salnikow and B. Wittmann-Liebold, Eur. J. Biochem. 244, 750 (1997). 12. T. Bergman, EXS 88, 133 (2000). 13. A.R. Dongré, J.L. Jones, A. Somogyi and V. Wysocki J. Am. Chem. Soc. 118, 8365 (1996). 14. S.G. Summerfield and S.J. Gaskell, Int. J. Mass Spectrom. Ion Proc. 165/166, 509 (1997). 15. K. Biemann, Methods in Enzymology 193, 455 (1990). 16. U. Bahr, M. Karas and R. Kellner, Rapid Commun. Mass Spectrom. 12, 1382 (1998). 17. M. Adamczyk, J.C. Gebler and J. Wu, Rapid Commun. Mass Spectrom. 13, 1413 (1999). 18. G. Tsaprailis, H. Nair, A. Somogyi, V.H. Wysocki, W. Zhong, J.H. Futrell, S.G. Summerfield and S. Gaskell, J. Am. Chem. Soc. 121, 5142 (1999).

Acknowledgement This work was supported by grants from CNRS (Programme Biologie Cellulaire), Association Française de

Received: 5 September 2000 Accepted: 24 March 2001 Web Publication: 5 October 2001