Proc. Nati. Acad. Sci. USA Vol. 82, pp. 3921-3924, June 1985 Neurobiology Identification of substance P precursor forms in human brain tissue (prohormones/tachykinins/hypothalamus/substantia nigra/caudate nucleus) FRED NYBERG, PIERRE LE GREvls, AND LARS TERENIUS Department of Pharmacology, University of Uppsala, S-751 24 Uppsala, Sweden Communicated by Tomas Hokfelt, January 28, 1985

ABSTRACT Substance P prohormones were identified in glycine (19). Enzymes involved in the processing of the the caudate nucleus, hypothalamus, and substantia nigra of precursors are incompletely known. human brain. A polypeptide fraction of acidic brain extracts The present paper reports the identification of substance P was fractionated on Sephadex G-50. The Iyophilized fractions precursors in human brain. The strategy for the experimental were sequentially treated with trypsin and a substance P- approach is based on the enzymatic generation in vitro of a degrading enzyme with strong preference toward the Phe7- unique peptide fragment (20, 21). Trypsin treatment of any Phe' and Phe8-Gly9 bonds. The released substance P(1-7) product of the preprotachykinins containing the substance P fragment was isolated by ion-exchange chromatography and sequence generates a fragment with the NH2-terminal region quantitated by a specific radioimmunoassay. Confirmation of identical with that of the native peptide. Further conversion the structure of the isolated radioimmunoassay-active frag- of this fragment with an endopeptidase capable of hydrolyz- ment was achieved by electrophoresis and HPLC. By using this ing substance P at the Phe7-Phe8 bond releases the substance enzymatic/radioimmunoassay procedure, two polypeptide P(1-7) sequence as a fragment that can be recovered by fractions of apparent M, 5000 and 15,000, respectively, were ion-exchange chromatography and quantitated by a specific identified. The latter component was the majior one of the two radioimmunoassay (RIA) (Fig. 1). This procedure was used but was estimated to account for only about 5% of total to identify substance P-containing polypeptide fractions in substance P radioimmunoassay activity. Because it is ofthe size human caudate nucleus, hypothalamus, and substantia nigra. predicted from the nucleotide sequences ofcDNA for substance A partial sequence of a protachykinin in bovine brain was P prohormones in bovine brain, the Mr 15,000 component may recently reported (22). represent the full-length prohormone. The primary structure of the undecapeptide substance P was MATERIALS AND METHODS determined in 1971 (1). This knowledge greatly facilitated Materials. Human caudate nucleus, hypothalamus, and studies on its putative role as neurotransmitter or substantia nigra were removed at autopsy and stored frozen at neuromodulator. Within the central nervous system, sub- -80'C before use. SP-Sephadex C-25, Sephadex G-50, and stance P has been localized in various nerve fibers (2-5) and Agarose C were obtained from Pharmacia. All reference pep- in cellular organelles from nerve endings (6-9). Substance P tides were from Peninsula Laboratories (San Carlos, CA) is released from nerve tissue within the central nervous except for the substance P(1-7), -(3-7), and -(1-8) fragments, system by depolarizing stimuli (7, 8, 10, 11) and can alter the which were prepared by J. M. Stewart (Department of Bio- activity of certain neurons when applied in their vicinity chemistry, University of Colorado, Denver). L-1-Tosylamido- (12-14). However, until recently very little was known about 2-phenylethyl chloromethyl ketone (TPCK)-treated trypsin and the biosynthesis of substance P. the protein standards used for column calibration in the gel- Substance P belongs to a family of structurally related filtration experiments were from Sigma or from KABI peptides, the tachykinins (15). Peptides in this family exhibit Diagnostica (Stockholm, Sweden; '2-I-labeled human prolac- biological activities similar to those of substance P and share tin). All other chemicals and solvents were of reagent grade. a common COOH-terminal sequence, Phe-Xaa-Gly-Leu- Enzyme Preparation. The enzyme activity capable of Met-NH2, the COOH-terminal residue of which is amidated. hydrolyzing substance P at the Phe7-Phe8 bond was purified The NH2-terminal sequence is unique for each peptide and from human cerebrospinal fluid by fractionation on DEAE- accounts for its specific action. In addition to substance P, Sepharose CL-6B, Sephadex G-100, and Sephacryl S-200 as only a few other members ofthe tachykinin family have been described (23). The cerebrospinal fluid specimens were identified in mammalian tissue (16, 17). The primary struc- obtained from a number of neurologic patients. The re- tures of two types of precursors of substance P from bovine covered enzyme (apparent Mr 40,000) was characterized as the use of cDNA an endopeptidase with high specificity for substance P, with brain were deduced by technology (18): no activity against, for instance, neurokinin A. Substance a-preprotachykinin contains a single substance P sequence P(1-7) and -(1-8) fragments are formed at a constant 3:2 ratio whereas f3-preprotachykinin contains not only substance P over time (23). The endopeptidase was obtained in a yield of but also a kassinin-like sequence, neurokinin A (substance about 1 ,ug/ml of cerebrospinal fluid and had an apparent Km K). In the precursors, each tachykinin sequence is followed of20 uM for the generation of the substance P(1-7) fragment. at the COOH-terminus by a glycine residue and is flanked by Recovery of Polypeptide Fractions from Brain Tissues. The pairs of basic amino acid residues, Arg-Arg or Lys-Arg (Fig. frozen brain tissues (5-12 g of caudate or hypothalamus, 1-2 1). The generation of substance P from its precursor probably g of substantia nigra) were thawed and homogenized in 1 M occurs in several steps leading to the liberation of a acetic acid (20 ml/g of tissue) at 90°C as described (24). The dodecapeptide, which would yield the amidated undecapep- extract collected after centrifugation was fractionated on a tide upon enzymatic conversion of the COOH-terminal Sephadex G-50 column (5 cm x 95 cm) equilibrated in 1 M acetic acid. Fractions of 20 ml were collected at a flow rate The publication costs of this article were defrayed in part by page charge of 120 ml/hr and lyophilized. payment. This article must therefore be hereby marked "advertisement" Enzymatic Degradation and Recovery of Radioimmunoassay- in accordance with 18 U.S.C. §1734 solely to indicate this fact. Active Fragments. The freeze-dried fractions were dissolved in 3921 Downloaded by guest on September 28, 2021 3922 Neurobiology: Nyberg et al. Proc. Natl. Acad. Sci. USA 82 (1985)

1 20 40 60 80 100 120 140

< < X 4<( At TIC / t

a-Preprotachykinin

44 0 CB o U) UOj0 < < Ji < 444ot - .._Xi

Putative signal peptide Substance P Neurokinin A P-Preprotachykinin

Prohormone 60 70 - -Arg-Arg-Pro-Lys-Pro-Gln-Gln-Phe-Phe-Gly-Leu-Met-Gly-Lys-Arg--

Trypsin treatment SP-Sephadex fractionation

Tryptic fragment

Arg-Pro-Lys-Pro-Gin-Gin-Phe-Phe- Gly-Leu-Met-Gly-Lys

Cleavage with specific enzyme I SP-Sephadex fractionation Substance P(1-7)

Arg-Pro-Lys-Pro-Gln-Gln-Phe

Radioimmunoassay FIG. 1. Schematic structure of a- and 1-preprotachykinin from bovine brain (18) and outline of the experimental strategy used to generate substance P(1-7). The substance P and neurokinin A units are indicated by the solid boxes. The hatched areas represent the putative signal peptide. Numbers indicate amino acid position, starting with the first residue of the signal peptide. Trypsin treatment releases a tridecapeptide, which then is isolated by ion-exchange chromatography (SP-Sephadex). Further cleavage by specific endopeptidase from cerebrospinal fluid liberates substance P(1-7). The heptapeptide fragment is isolated by a second SP-Sephadex chromatography and quantified by radio- immunoassay.

300 pl of 0.4 M N-ethylmorpholine acetate, pH 8.2. After the HPLC. Samples were electrophoresed in 1.0 cm x 65-cm addition of50 A.l oftrypsin solution (2 mg/ml in the same buffer) columns of agarose suspension, according to a published to each tube, the tubes were incubated for 8 hr at 37TC. The procedure (26). Electrophoresis was in acidic pH at 1200 V reaction was terminated by boiling for 2-3 min before dilution and 15 mA for 5 hr. Aliquots (0.5 ml) ofthe suspensions were with 4 ml of0.1 M pyridine formate, pH 4.5. Aftercentrifugation eluted and subjected to radioimmunoassay. at 10,000 x g for 5 min, each sample was fractionated on HPLC was performed in a Hewlett-Packard 1084B liquid SP-Sephadex C-25 according to a procedure described previ- chromatograph with a Waters ,uBondapak C18 column (4.5 x ously (25). The ion exchanger had been equilibrated at pH 4.5 250 mm). Samples were eluted with a 40-min linear 15-30% in the above pyridine formate buffer and elution was with 1.6 M gradient ofmethanol in 0.04% trifluoroacetic acid followed by pyridine formate at the same pH. The eluate obtained from each 30% methanol/0.04% trifluoroacetic acid. Fractions of 1 ml sample was evaporated (Savant Vac concentrator) and recon- were collected at a flow rate of 1 ml/min, evaporated, and stituted in 50 jul of0.2 M Tris Cl, pH 7.8. Addition of 50 A.l (0.5 analyzed by radioimmunoassay. The HPLC column and the Ag) of the cerebrospinal fluid enzyme was followed by a 2-hr electrophoresis system were calibrated by using substance P incubation at 370C before the reaction was terminated as before. and its 1-7, 3-7, 3-11, and 5-11 fragments as markers. The SP-Sephadex fractionation procedure was repeated except Radioimmunoassay. The radioimmunoassay for substance that the eluate was collected at a lower buffer concentration (0.8 P(1-7) was described previously (24). 3H-labeled substance M). Fractions were dried and then were analyzed by radio- P(1-7) was used as tracer with antiserum that had been raised immunoassay. in rabbits against the peptide-thyroglobulin conjugate. The Identification of Substance P(1-7) by Electrophoresis and antiserum gave about 1% crossreactivity with substance Downloaded by guest on September 28, 2021 Neurobiology: Nyberg et al. Proc. Natl. Acad. Sci. USA 82 (1985) 3923 P(1-8), whereas its crossreactivity with substance P and precursor represents only about 5% of the total substance other substance P fragments was <0.2%. P(1-7)-containing peptides, based on radioimmunoassay ac- tivity. Experiments performed with synthetic substance P RESULTS showed that the cerebrospinal fluid enzyme can convert, at most, 40% of the peptide to its 1-7 fragment. The yield from Fig. 2A illustrates the elution pattern from a Sephadex G-50 picomole amounts of synthetic substance P over the whole column of substance P(1-7)-containing peptides in a crude degradation procedure, including trypsin treatment and SP- hypothalamus extract. There are two fractions ofapparent Mr Sephadex fractionation, did not exceed 10%. With nanomole 15,000 and 5000, respectively, whereas the bulk of substance amounts, the yield improved to about 25% for both substance P(1-7) activity was recovered in the low Mr region of au- P and the substance P-glycine derivative. thentic substance P and its 1-7 fragment. Repetitive gel- Results from HPLC analysis of the products reacting in the filtration runs of hypothalamic material yielded similar dis- substance P(1-7) radioimmunoassay are given in Fig. 3. The tributions, with some quantitative variations among the activity derived from the Mr 15,000 component from the various individual patients. In a separate experiment, where the brain areas was eluted at the same position as the synthetic extract was fractionated on a SP-Sephadex column prior to heptapeptide. HPLC of the activity liberated by enzyme treat- Sephadex G-50 chromatography, an additional substance ment of the Mr 5000 species gave the same result. P(1-7)-containing component of Mr -2000 was distinguished Analytical electrophoresis also confirmed the identity of in addition to the Mr 5000 species (Fig. 2B). Sephadex G-50 the substance P(1-7) fragment liberated from the Mr 15,000 chromatography ofextracts from human caudate nucleus and fraction from hypothalamus (Fig. 4). The same observation substantia nigra yielded elution profiles in conformity with was made for the corresponding material recovered from those shown in Fig. 2A. The substance P(1-7) activity caudate nucleus extracts. recovered from the higher Mr fractions was always much lower than that from the low Mr region. For instance, the DISCUSSION experiment shown in Fig. 2A indicates that the Mr 15,000 To identify all possible substance P peptides and particularly larger precursors, we developed a two-step enzymatic pro- cedure with isolation by ion-exchange chromatography both a. cI) of an intermediate tridecapeptide, formed by trypsin treat- 6.0 ment, and of the final product, substance P(1-7), formed by A 11 1 4.0- I RR~~~~~G_ 2.

120 n ALPr A 80- 0 10 0.) '.2 R 40 4-E

.1 20 40 60 80 100 0 B I i.6- o IL CI- C,)0) 0) '5 100 O c0 1 11 1 CL-C .4 o0 1200- f 0 B 10 o 0..2 c4M CDC

0 20 40 60 80 100 Fraction FIG. 2. The distribution of substance P(1-7)-containing fractions after Sephadex G-50 chromatography of human hypothalamic ex- tracts. Fractions were and sequentially treated with lyophilized 20 40 60 trypsin and an endopeptidase before being analyzed for substance 0 P(1-7) radioimmunoactivity. The extracts, originating from 10 g (A) Fraction and 12 g of tissue (B), were applied without (A) or after (B) prior fractionation on SP-Sephadex C-25. In B, only fractions 53-77 were FIG. 3. HPLC analysis of the substance P(1-7)-like activity assayed. The column (5 x 95 cm) was calibrated with the following liberated from the Mr 5000 (A) and 15,000 (B) fractions recovered protein and peptide standards: bovine serum albumin (Mr 68,000), from human hypothalamus and from combined MA 5000 and 15,000 human prolactin (Mr 22,000), bovine pancreas RNase (Mr 14000), material from substantia nigra (C). Fractions of the eluate were ,8-endorphin (Mr 3500), dynorphin A (Mr 2100), substance P (Mr analyzed by radioimmunoassay. In calibration runs, the substance P 1300), and substance P(1-7) (Mr 900). P. substance P. standards were eluted as indicated by the arrows. Downloaded by guest on September 28, 2021 3924 Neurobiology: Nyberg et al. Proc. Natl. Acad. Sci. USA 82 (1985)

_ pmol/g of tissue, represents about 5% of all activity gener- I ated from the tissue. However, considering the compara- tively modest yield of the enzyme degradation procedure, it A ~ J CO + is likely that the true concentration is at least 10 times higher. This would also be true for the activity that was coeluted with 1.6- substance P (Fig. 2A). In fact, the substance P(1-7) activity recovered in the low molecular weight region by the enzy- 0) matic procedure only represented about a tenth of that c determined by direct radioimmunoassay of substance P. 0 a Postmortem changes may have affected the relative concen- 8 trations of the precursor forms. However, there was no CD -0 ,o significant difference between tissues obtained 10 and 30 hr 5 . after death. E The method described here should allow studies of the Iv- regulation of biosynthesis of the substance P precursor and r- 1 the mechanisms of precusor processing. CL 0- 1.16XB D This work was supported by the National Institute on Drug Addiction (DAC:1502) and the Swedish Medical Research Council coU)0 (04-03760). C: B- .0 1. Chang, M. M., Leeman, S. E. & Niall, H. D. (1971) Nature (London) New Biol. 232, 86-87. 2. Hbkfelt, T., Kellerth, J. O., Nilsson, C. & Pernow, B. (1975) Science 190, 889-890. 3. Hokfelt, T., Johansson, O., Ljungdahl, A., Lundberg, J. M. & 0 20 40 60 Schultzberg, M. (1980) Nature (London) 284, 515-521. Fraction 4. Mroz, E. A., Brownstein, M. J. & Leeman, S. E. (1976) Brain Res. 113, 597-599. FIG. 4. Electrophoresis in agarose suspension (0.18%, wt/vol) of 5. Cuello, A. C. & Kanazawa, I. (1978) J. Comp. Neurol. 178, substance P(1-7)-like immunoactivity released from the M, 15,000 129-156. component recovered from human hypothalamus (A) and caudate 6. Duffy, M. J., Mulhall, D. & Powell, D. (1975) J. Neurochem. nucleus (B). Eluted fractions were screened for substance P(1-7) 25, 305-307. activity by radioimmunoassay. The column was calibrated with 7. Schenker, C., Mroz, E. A. & Leeman, S. E. (1976) Nature substance P standards as indicated by the arrows. (London) 264, 790-792. 8. Lembeck, F., Mayer, N. & Schindler, G. (1977) Naunyn- treatment with a highly selective endopeptidase. The con- Schmiedeberg's Arch. Pharmacol. 301, 17-22. centration of generated substance P(1-7) was by 9. Palay, V. C. & Palay, S. L. (1977) Proc. Nat!. Acad. Sci. USA quantified 74, 3597-3601. radioimmunoassay, and its structural identity was confirmed 10. Iversen, L. L., Jessell, T. & Kanazawa, I. (1976) Nature by two independent analytical techniques. In control experi- (London) 264, 82-83. ments, trypsin treatment did not affect substance P or the 1-7 11. Otsuka, M. & Konishi, S. (1976) Nature (London) 264, 83-84. fragment. Since the procedure has a series of critical steps, 12. Konishi, S. & Otsuka, M. (1974) Nature (London) 252, it should have the desired specificity. The substance P 734-735. endopeptidase was chosen because of its specificity and 13. Krnjevic, K. (1977) in Substance P, eds. von Euler, U. S. & potency. Pernow, B. (Raven, New York), pp. 217-230. The size and structure of the bovine prohormones for 14. Anderson, R. K., Lund, J. P. & Ruil, E. (1978) Can. J. Physiol. substance P have been predicted from cDNA sequences (18). Pharmacol. 56, 216-222. human 15. Erspamer, V. (1981) Trends Neurosci. 4, 267-269. Corresponding information for the prohormones is 16. Lazarus, L. H., Linnoila, R. I., Hernandez, 0. & DiAgustine, not available. Assuming cross-species structure conserva- R. P. (1980) Nature (London) 287, 555-558. tion, the Mr 15,000 fraction may represent the full sequence 17. Minamino, N., Kangawa, K., Fukuda, A. & Matsuo, H. (1984) of a- or (-protachykinin, whereas the Mr 2000 and 5000 Neuropeptides 4, 157-166. fractions may represent shorter intermediates in the process- 18. Nawa, H., Hirose, T., Takashima, H., Inayama, S. & ing of the precursor containing the substance P unit. As in Nakanishi, S. (1983) Nature (London) 306, 32-36. other peptide precursors, the proteolytic processing probably 19. Harmar, A., Schofield, J. G. & Keen, P. (1981) Neuroscience occurs at pairs of basic amino acid residues that flank the 6, 1917-1922. mature hormone sequence. If this is the case and if the Mr 20. Wang, S.-S. & Carpenter, F. H. (1965) J. Biol. Chem. 240, 5000 found in the human brain is an 1619-1625. species incomplete 21. Lewis, R. V., Stein, S., Gerber, L. D., Rubinstein, M. & cleavage product(s) of preprotachykinin, then that species is Udenfriend, S. (1978) Proc. Nat!. Acad. Sci. USA 75, likely to consist of substance P plus COOH-terminal flanking 4021-4023. sequence (e.g., corresponding to the residues 58-108 in the 22. Tatemoto, K. & Mutt, V. (1984) Digestive Dis. Sci. New Ser. bovine prohormones; see Fig. 1). It should be noted that our 29, 88 (abstr. 87). molecular weight estimates are tentative and may be over- 23. Nyberg, F., Le Greves, P., Sundqvist, C. & Terenius, L. estimated by as much as 20% in the region above Mr 10,000. (1984) Biochem. Biophys. Res. Commun. 125, 244-250. The radioimmunoassay analysis indicates that the pre- 24. Stewart, J. M., Hall, M., Harkins, J., Frederickson, R., cursor fractions identified in human brain represent only a Terenius, L., Hokfelt, T. & Krivoy, W. (1982) Peptides 3, small fraction of total substance P-related peptides. This is 851-857. 25. Bergstrom, L., Christensson, I., Folkesson, R., Stenstrom, B. compatible with studies of labeled amino acid incorporation & Terenius, L. (1983) Life Sci. 33, 1613-1619. into substance P in vivo that indicate a rapid turnover of 26. Nyberg, F. & Terenius, L. (1983) Anal. Biochem. 134, precursor. Four hours after a pulse of labeled amino acids, 245-253. radioactivity is recovered in the mature peptide (27). The 27. Krause, J. E., Reiner, A. J., Advis, J. P. & McKelvy, J. F. heptapeptide released from the Mr 15,000 material, 0.5 (1984) J. Neurosci. 4, 775-785. Downloaded by guest on September 28, 2021