RAPID COMMUNICATIONS IN MASS SPECTROMETRY, VOL. 6,637-640 (1992)

C-Terminal Sequencing of Peptides Using Electrospray Ionization Mass Spectrometry Kenneth J. Rosnack and Justin G. Stroh* Central Research Division, Pfizer Inc, Groton, CT 06340, USA

A low-flow reactor is described for the on-line monitoring of peptides digested with carboxypeptidase P by electrospray ionization. Two peptides were analyzed using this technique: glucagon (average MW 3482.8 Da), and apomyoglobin (average MW 16 951.5). Both peptides gave interpretable results. The first 19 amino acids of glucagon were successfully sequenced. Apomyoglobin yielded sequence information to the 30th amino acid with some gaps. At 300 nL/min, 50% of the first 30 amino acids were sequenced and at 1 pL/min, 67% of the first 30 amino acids were observed.

Electrospray ionization mass spectrometry has been used to sequence a variety of pep tide^.'-^ Numerous methods have been employed in this regard with varying degrees of success. One of the more common approaches involves determining the molecular weight of the intact peptide followed by liquid chromatography combined with mass spectrometry (LC/MS) and LC/MS/MS analysis of a tryptic digest of the intact peptide. Trypsin (which cleaves at the basic amino acids lysine and arginine) is often chosen first because it is highly specific, can be easily obtained in sequencing grade purity, and most importantly provides a dibasic peptide for every enzymatic fragment, save the C-terminal fragment. Dibasic fragments tend to ionize well by electrospray ionization as the doubly charged species, hence sensitivity is often high. However, the C-terminal tryptic fragment is usually monobasic (unless the C-terminal residue of the intact protein is lysine or arginine) and is often unseen in the electrospray LC/MS experiment. One possible method for overcoming this difficulty is to C-terminally sequence the intact protein until an overlap with the observed tryptic fragments can be found. Two basic approaches are currently used for C-terminal sequencing: chemical and enzymatic cleavage. Chemical cleavage has recently been used in conjunction with mass spectrometry6.' to produce some very useful results. On the other hand, C-terminal sequencing using enzymes has also produced useable data. With enzymatic digestion, sequencing is usually performed with one or more of the carboxypeptidases, which selectively cleave the C-terminal amino acid from the peptide, and the released amino acids are analyzed. However, sequence analysis based on amino acid release is difficult due to non-uniform rates of digestion and the possibility of repetitive amino acid residues. On the other hand, numerous mass spectral techniques have been used with carboxypeptidase enzymes to monitor the course of digestion including fast-atom bombardment (FAB),8-'2 thermospray,*3-17 plasma desorption," laser desorption,lg field desorption,*".*' and most recently, electrospray.22With mass spectrometry, the molecular weight of the peptide that remains after sequential digestion is monitored. Monitoring the molecular weights of the residual peptides circumvents the problem of repetitive amino acid residues. Also, a mixture of peptides produced by nonlinear kinetics is directly observable by mass spectrom09.51-4198/92/110637-04 $07.00 01992 by John Wiley & Sons, Ltd

etry since each peptide in the mixture has a different mass. The standard method of analysis with carboxypeptidase enzymes is to sample the reaction mixture at different time points by removing an aliquot, quenching the enzymatic reaction and analyzing. This method works quite well as long as the rates of cleavage are uniform. However, it is known that the rate of cleavage is related to the primary, secondary and tertiary structure of the target and may differ by orders of magnitude in rate. For an unknown, this may be difficult. One solution to this problem is to continuously monitor the reaction with a flowing stream, for which electrospray is particularly well suited. We have constructed and tested an on-line reactor for carboxypeptidase digestions, which is the subject of this report. EXPERIMENTAL Experiments were performed on a Finnigan TSQ-700 triple quadrupole mass spectrometer (San Jose, CA, USA) using an electrospray source (Analytica of Branford, Branford, CT, USA). Data were collected in the profile mode (average=4) over a mass range of 400-1800 Da or 600-2200 Da with a scan time of 10 s. The electrospray source was operated at 3500 V (3060 pA) with a curtain gas temperature of 200-250 "C. Processed data were averaged over 10-20 scans, smoothed using a 7 point smoothing routine, d e c o n ~ o l u t e d ,and ~ ~ manually centroided. Glucagon (special quality) and apomyoglobin (horse skeletal muscle) standards were obtained from Boehringer Mannheim (Indianapolis, IN, USA), and Sigma (St Louis, MO, USA), respectively, and were used as received. The carboxypeptidase enzymes (Y and P-sequencing grade, A and B-standard grade) were obtained from Boehringer Mannheim and were used as received. All solvents were obtained from Baker (Phillipsburg, NJ, USA). Fused silica tubing (Polymicro Technologies Incorporated, Phoenix, AZ, USA) and 1 / 1 6 x 0.01'' high pressure Teflon (VWR Scientific, Boston, MA, USA) were used to construct the reactor. A VWR Model 1225 constant temperature bath was used to control the temperature of the reactor. To minimize dead volume, the inlet and outlet of the reactor were constructed from a 50 cm and 65 cm length of 50 pm I D fused silica, respectively. The reactor was constructed of 75 cm of 320 pm fused silica tubing Received 14 Auguri 1992 Accepted (revised) 21 September I992

C-TERMINAL SEQUENCING O F PEPTIDES USING ESP-MS

638 TO Mass Spectrometer L

Teflon Tubing 0.010" I.D.

50 L ID Fused Silica Tubing 650 Inmi

-

v Tubing

-i

500 mm

l 80 o0l

(digestion of apomyoglobin was also performed using a 275 cm reactor, 220 pL volume) and was joined to the inlet/outlet using 0.0625" X 0.01'' high pressure Teflon tubing. The inlet and outlet were connected to a 100 yL Hamilton Syringe and the electrospray source respectively using the same high pressure Teflon tubing. The reactor was maintained at 37 "C in a water bath with the Teflon connectors above the water level. See Fig. 1 for the reactor schematic diagram. Glucagon (30 pmol/ pL) and apomyoglobin (25 pmol/ pL) were each dissolved in 0.1% acetic acid. Individual enzymes (P or Y) and mixtures of the enzymes (P, Y or P, B, A ) were used to digest the samples (the substrate:enzyme ratio was 100:1to 50: 1for each enzyme used). The digest mixture was infused into the 60pL reactor (a 220 pL reactor was also used for the digestion of apomyoglobin). Once the reactor was full, a 100 pL Hamilton syringe (Hamilton Company, Reno, NV, USA) was loaded with 0.1% acetic acid to push the digest mixture through the reactor into the electrospray source at 300 nL/min (digestion of apomyoglobin was also performed at 1 pL/min and required refilling of the syringe). A 500 pL Unimetrics syringe (Unimetrics, Shorewood, IL, USA) containing 0.1% trifluoroacetic acid in 2-methoxyethanol provided a sheath liquid at a flow rate of 1.5 pL/min. Digestion of apomyoglobin was also performed using the above sheath liquid in a 100 pL Hamilton syringe flowing at 1pL/min (the syringe required refilling during the digestion). The digest was monitored continuously on-line for up to 3 h (apomyoglobin was also monitored for up to 3.75 h). RESULTS AND DISCUSSION Our initial experiments focused on choosing the best carboxypeptidase enzyme or mixture of carboxypeptidase enzymes for use with electrospray. The four carboxypeptidase enzymes tested were carboxypeptidase A , B, P, and Y. A mixture of carboxypeptidase A , B, and P (at the optimum pH for carboxypeptidase A) was tested in electrospray experiments, but produced signal suppression. Perhaps this is due to the fact that carboxypeptidase A is a suspension in water that is saturated with toluene and both carboxypeptidase A and B have an optimum pH of 7, which is higher than desired for the electrospray experiment. These enzymes were not tested further. Carboxypeptidase Y produced acceptable results (optimum pH = 6). However, carboxypeptidase P (optimum p H = 5) was 30 times more sensitive than carboxypeptidase Y, presumably because of the lower optimum pH. In our hands, a mixture of 1:l carboxypeptidases P and Y was twelve times less sensitive than straight carboxypepti-

2dOO

2600

3600

31'00

32b0

3300

34b0

3500

m/z Figure 2. Deconvoluted spectrum of glucagon. Concentration at 30 pmollpL. Flow rate at 300nL/min. Average centroided mass and corresponding amino acid loss are indicated.

dase P. Therefore, in the following experiments carboxypeptidase P was used as the enzyme of choice. Having chosen carboxypeptidase P as the enzyme, an on-line reactor was used to monitor a complete digestion of glucagon, a 29 amino acid peptide. A deconvoluted mass spectrum taken 25 min into the run is shown in Fig. 2. The spectrum shows the loss of T , N, M, L, and W which are the first five amino acids in the sequence. A peak at 3265.4Da, corresponding to cleavage at TN, is small and broad resulting in poor mass accuracy. All other ions observed in this experiment could be measured accurately enough to define the correct amino acid sequence. Poor mass accuracy may lead to incorrect assignments in certain cases. When a sequence contains amino acids that are close in mass to other amino acids, higher accuracy is required. For instance, to distinguish glutamine from glutamic acid (mass difference (A) = 1Da) in a peptide of molecular weight 3483 Da requires a mass measurement accuracy of 0.03%. On the other hand, to distinguish alanine from serine (A = 16 Da) for the same peptide

Table 1. Glucagon. Column 1, sequence position starting from the C-terminus; column 2, amino acid; column 3, found (+)=yes; column 4, calculated average mass of remaining peptide Position

1 2 3 4 5 6 7

Amino acid

T N M L W

Q V

8

F

9 10 11 12 13 14 15 16 17 18 19

D

Q A R R

s

D L Y K S

Found

+ + + + + + + + + + + + + + + + + + +

Mass (Da)

3482.8 3381.7 3267.6 3136.4 3023.2 2837.0 2708.9 2609.8 2462.6 2347.5 2219.4 2148.3 1992.1 1835.9 1748.8 1633.7 1520.6 1357.4 1229.2

C-TERMINAL SEQUENCING OF PEPTIDES USING ESP-MS

x

N

OD

In

100,

I

I

I

I

1

I

I

I

I

I

15600 15800 16000 16200 16400 16600 16800 17000

m/z Figure 3. Deconvoluted spectrum of apomyoglobin. Concentration at 25 pmollpL. Flow rate at 300 nllmin. Average centroided mass and corresponding amino acid(s) loss are indicated.

only requires a mass measurement accuracy of 0.5%. Of course, using this technique, neither leucine nor isoleucine (113 Da) can be distinguished nor can glutamine or lysine (128 Da). The reaction was stopped after 216 min at which time the sequence had been extended to include 19 amino

Table 2. Apomyoglobin. Column 1, sequence position starting from the C-terminus;column 2, amino acid; column 3, found (+)=yes, (-)=no for experiment at flow rate =0.3 yLlmin; column 4, found (+) =yes, (-) = no for experiment at flow rate= 1.0 pL/min; column 5 , calculated average mass of remaining peptide Position

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Amino acid

G

Q F G L E K Y K A A I D N R

F L E L A K T M A G

Q A D A G

Found (1)

+ + +

-

+ + + + + -

-

Found (2)

+ + +

-

+ + + + + + +

-

-

-

-

-

-

+ + + + + + +

+ + + + + + + + + +

Mass (Da)

16 951.5 16 894.5 16 766.3 16 619.2 16 562.1 16 449.0 16 319.8 16 191.7 16 028.5 15 900.3 15 829.2 15 758.2 I5 645.0 15 529.9 15 415.8 15 259.6 15 112.4 14 999.3 14 870.2 14 757.0 14 685.9 14 557.7 14 456.6 14 325.4 14 254.4 14 197.3 14 069.2 13 998.1 13 883.0 13811.9

639

acids (see Table 1). A total of 1.9 nmol of glucagon was consumed in this experiment. A potential application of this technique is C-terminal sequencing of proteins. We tested three proteins using this technique: cytochrome c , carbonic anhydrase, and apomyoglobin. Of these three, only apomyoglobin digested with carboxypeptidase P, suggesting that this technique will have limited applicability to proteins. No attempt was made to denature the proteins prior to analysis whereas denaturation may have allowed the carboxypeptidase reaction to proceed. A deconvoluted spectrum of digested apomyoglobin taken at 42min into the run is found in Fig. 3. The results of two experiments using apomyoglobin are given in Table 2, showing that after 3 h of digestion, 30 amino acids had been cleaved with 50% of the sequence being observed. The data presented in Table 2 indicate that the first amino acid after every glycine residue was not observed (positions 2, 5, and 26). The cleavage of glycine (and serine) residues is known to be slow, and if the next cleavage is very fast it should not be observed due to a low steady state concentration. Also, there are two large gaps in the sequence (positions 5-9 and 15-22). At a higher flow rate (1 pL/ min), more of the sequence was observed (the gap from position 15-21 was reduced to a gap between positions 17-19), suggesting that cleavage was occurring at most if not all of the amino acids, but that the concentration of these peptides was low. The data presented here indicate that sequencing of peptides with carboxypeptidase P and electrospray ionization using an on-line reactor is feasible. Glucagon gave sequence ions for the first 19 amino acids; however, apomyoglobin gave only 50% of the first 30 amino acids. The data suggest that sequence information is highly dependent on the proteolytic reactivity of the enzyme and substrate. REFERENCES 1. C. K. Meng, C. N. McEwen and B. S. Larsen, Rapid Commun. Mass Specirom., 4, 151 (1990). 2. P. R. Griffin, J . A. Coffman, L. E . Hood and J. R. Yates, 111, Ini. J. Mass Spectrom. Ion Processes, 111, 131 (1991). 3. J. B. Smith, G. Thevenon-Emeric, D. L. Smith and B. N. Green, Anal. Biochem., 191, 118 (1991). 4. R. A. Henderson, H. Michel, K. Sakaguchi, J . Shabanowitz, E . Appella, D. F. Hunt and V. H . Engelhard, Science, 255, 1264 (1992). 5. D. F. Hunt, M. Hanspeter, T. A. Dickinson, J . Shabanowitz, A. L. Cox, K. Sakaguchi, E. Appella, H. M. Grey and A. Sette, Science, 256, 1817 (1992). 6. A. Tsugita, K. Takamoto and K. Satake, Chem. Letts., 235 (1992). 7. R. Aebersold, E. J. Bures, M. Namchuk, M. H. Goghari, B. Shushan and T. C. Covey, Protein Sci., 1, 494 (1992). 8. C. V. Bradley, D. H. Williams and M. R. Hanley, Biochem. Biophys. Res. Commun., 104, 1223 (1952). 9. R. Self and A. Parente, Biomed. Mass Spectrom., 10.78 (1983). 10. L. A. Smith and R. M. Caprioli, Biomed. Mass Spectrom., 10,98 (1983). 11. R. M. Caprioli and T. Fan, Anal. Biochem., 154, 596 (1986). 12. R. M. Wagner and B. A. Fraser, Biomed. Enuiron. Mass Specfrom., 14, 235 (1987). 13. D . Pilosof, H.-Y. Kim, M. L. Vestal and D. F. Dyckes, Biomed. Mass Specirom., 11, 403 (1984). 14. H.-Y. Kim, D. Pilosof, D. F. Dyckes and M. L. Vestal, J. Am. Chem. Soc., 106, 7304 (1984). 15. K. Stachowiak, C. Wilder, M. L. Vestal and D. F. Dyckes, J . Am. Chem. Soc., 110, 1758(1988).

640

C-TERMINAL SEQUENCING OF PEPTIDES USING ESP-MS

16. K. Stachowiak and D. F. Dyckes, Peptide Res., 2 , 267 (1989). 17. R. D. Voyksner, D. C. Chem and H. E . Swaisgood, Anal. Biochem., 188, 72 (1990). 18. K. Klarskov, K. Breddam and P. Roepstorff, Anal. Biochem., 180, 28 (1989). 19. M. Schaer, K. 0. Boernsen, E . Gassmann and H. M. Widmer, Chimia, 45, 123 (1991). 20. A. Tsugita, R. Van Den Broek and M. Przybylski, FEBS Lett., 137, 19 (1982).

21. Y. M. Hong, T. Takao, S. Aimoto and Y. Shimonishi, Biomed. Mass Spectrom., 10, 450 (1983). 22. K. L. Duffin and C. E. Smith, Proceedings ofthe 40th Conference on Mass Spectrometry and Allied Topics, Washington, DC, ASMS, East Lansing (1992). 23. J. Zhou and I. Jardine, Proceedings of the 38th Conference on Mass Spectrometry and Allied Topics, Tucson, AZ, ASMS, East Lansing (1990).