Biochem. J. (1984) 218, 857-862 857
Printed in Greqt Britain
D-Lysergic acid-activating enzyme from the ergot fungus Claviceps purpurea
Ullrich KELLER*, Rainer ZOCHER, Ulrike KRENGEL and Horst KLEINKAUF
Technische Universitat Berlin, Institut fur Biochemie und Molekulare Biologie, Franklinstraf,e 29, D-1000
Berlin 10, Germany
(Received 5 October 1983/Accepted 23 November 1983)
A D-lysergic acid-activating enzyme from the ergot fungus Claviceps purpurea was
purified about 145-fold. The enzyme was able to catalyse both the D-lysergic acid-
dependent ATP-pyrophosphate exchange and the formation of ATP from n-lysergic
acid adenylate and pyrophosphate. Both reactions were also catalysed to a decreased
but significant extent with respect to dihydrolysergic acid. The molecular mass of the
enzyme was estimated to lie between 135 and 140kDa. The involvement of the
enzyme in the biosynthesis of ergot peptide alkaloids is discussed.
Ergot peptides represent a class of peptide
derivatives of n-lysergic acid (Fig. 1). They are pro-
duced by the ergot fungus Claviceps purpurea. Bio-
synthetic studies in vivo revealed that the natural
amino acids are the precursors for the amino acids
present in the peptide chain (Gr6ger, 1975). In the
case of the n-lysergic acid moiety it was previously
demonstrated that D-lysergic acid, when added to
protoplasts or mycelium of Claviceps purpurea
A.T.C.C. 20102, stimulates ergot-peptide syn-
thesis significantly (Keller et al., 1980). In addi-
tion, feeding this organism with dihydrolysergic Fig. 1. Structural formula of ergot peptides
acid results in the formation of the corresponding Ergotamine: R= -CH3, R2 = -CH2-C6H5; ergo-
dihydro ergot peptides (Anderson et al., 1979). i> cryptine: R1 =-CH(CH3)2, R2 =-CH(CH3)-
Lysergic acid seems to be a free intermediate in the C2H5; ergosine: R1 = CH3, R2 =-CH2-
process of lysergyl-peptide formation, so an en- CH(CH3)2; ergocornine: R1 = CH(CH3)2, R2 =
zyme must be present in the cells of C. purpurea -CH(CH3)2-
catalysing the activation of D-lysergic acid before
peptide-bond formation. Such an activation is
likely to proceed via adenylate formation, as has
been described for numerous non-ribosomally Nuclear Corp. [a-32P]Adenosine triphosph4te tri-
formed peptides (Kleinkauf, 1979). In the present ethylammonium salt (39Ci/mmol) was from The
paper we describe the purification and character- Radiochemical Centre, Amersham, Bucks., U.K.
ization of a D-lysergic acid-activating enzyme from Agroclavine, chanoclavine, elymoclavine,
Claviceps purpureoa strain 1029. dihydrolysergic acid, dihydrolysergic acid amide
and D-lysergic acid methyl ester were kindly
Materials and methods provided by Dr. H. Kobel, Sandoz AG, Basel,
Chemicals and radioisotopes Switzerland. Ergotamine tartrate was from Sigma.
All other chemicals were of the highest purity
Sodium [32P]pyrophosphate (2.9Ci/mmol) and commercially available.
[U-14C]adenosine triphosphate tetrasodium salt
(569Ci/mol) were obtained from New England Syntheses
Abbreviations used: PMSF, phenylmethanesulphonyl D-Lysergic acid was prepared from ergotamine
fluoride; PEI-, polyethyleneimine-. tartrate as described by Stoll et al. (1943). The
To whom correspondence and requests for reprints adenylates of n-lysergic acid and dihydrolysergic
should be addressed. acid were prepared essentially by the method of
858 U. Keller, R. Zocher, U. Krengel and H. Kleinkauf
Berg (1958) with several slight modifications. The reaction mixtures were subjected to t.l.c. on PEI-
synthesis was performed at one-twentieth of its cellulose with solvent system III. Buffer was buffer
original scale under an atmosphere of N2. The B.
product was either kept as a solid or as a solution in
2-methoxyethanol (Moldave et al., 1959) at A TP formation from adenylate
- 20°C. The adenylates were identified by t.l.c. on
silica-gel plates at 4°C using solvent system I Adenylate (50-lOOpg) was dissolved in 50pl of
(butanol/acetic acid/water, 4:1:1 by vol.) or buffer B. After addition of 2 x 106 d.p.m. of [32p]-
solvent system II (ethanol/water, 4 :1, v/v) (Jaku- pyrophosphate, 1 pmol of MgCl2 and 120pl of
bowski et al., 1977; Rucman, 1976). The zones enzyme, the volume was adjusted to 2001p and the
corresponding to the adenylates were recognized mixture incubated at 24°C for 30min. A 601
either by fluorescence in the far u.v. (D-lysergic portion of the reaction mixture was added to 1 ml
acid) or after spraying with van Urk's reagent of Norit A charcoal suspension (Lee & Lipmann,
(Hofmann, 1964) (dihydrolysergic acid). The D- 1975). After several washings with distilled water,
lysergic acid and dihydrolysergic acid adenylates radioactive ATP was eluted from the charcoal with
were identified by scraping off the newly formed 10% (v/v) pyridine. The extract was evaporated to
bands from the silica-gel layer and eluting the dryness and the residue chromatographed on PEI-
compounds with water. After being left at room cellulose with solvent system IV.
temperature for 30min, the aqueous phases were
evaporated to dryness at 30°C. The residues were
taken up in a minute volume of water and applied Enzyme purification
to PEI-impregnated cellulose sheets (Macherey, All operations were carried out at 0-4°C. Buffer
Nagel and Co., Duren, Germany) and chromato- A contained 5 mM-sodium phosphate, pH 6, 5 mM-
graphed by using solvent system III (1.2M-LiCI) or dithioerythritol and 1 mM-PMSF. Buffer B con-
IV (2M-formic acid/l M-LiCl). The presence of tained 50mM-potassium phosphate, pH 6.8, 5mM-
AMP (detected by its fluorescence) revealed the dithioerythritol and 1 mM-PMSF.
nature of the compounds as nucleotides of D- Step 1. Freeze-dried mycelium from a 5-7-day-
lysergic acid or dihydrolysergic acid, since the RF old culture of Claviceps purpurea (6.5 g) was ground
values of the newly formed compounds were in a mortar for several minutes. The resultant fine-
distinctly higher than that of AMP and lower than powder-like material was stirred with 200ml of
that of the two ergolines. The RF values for the two buffer A for 1 h. The suspension was then
adenylates were similar (0.2-0.3 in solvent system centrifuged at 12000g for 30min. The supernatant
I, 0.3-0.4 in solvent system II). They were was applied on to a DEAE-cellulose column (DE
extremely labile at room temperature and decom- 52, Whatman; 2cm x 14cm). After washing with
posed slowly during chromatography, which made 400ml of buffer A, crude enzyme (25-30ml) was
it difficult to determine the RF value exactly. T.l.c. eluted with buffer A containing KC1 to give a
revealed that aqueous solutions of the adenylates conductivity of 7mS.
quickly prepared before enzymic tests contained Step 2. Crude enzyme (4ml) was fractionated on
30-50% adenylate. an Ultrogel AcA 34 column (1.6cm x 100 cm)
equilibrated with buffer B. Fractions (3.3 ml) were
Enzyme assay collected and assayed for the presence of D-lysergic
The ATP-pyrophosphate exchange was acid-activating enzyme by means of the ATP-PPi
measured by a procedure modified from that exchange reaction. This separation is shown in
described by Lee & Lipmann (1975). D-Lysergic Fig. 2. D-Lysergic acid-dependent ATP-PP1-ex-
acid or another ergoline (0.5mM), 5mM-MgCl2, change activity was found in fractions 26-34.
2.5mM-ATP, 0.25mM-sodium pyrophosphate (un- Normally activity that was not dependent on D-
less stated otherwise), 2x lO5d.p.m. of [32P]_ lysergic acid was less than 5% of the total.
pyrophosphate and 1001il of enzyme were allowed Fractions 27-33 were pooled.
to react together in a total volume of 20041 for Step 3. The pooled Ultrogel-AcA-34 fractions
1Omin at 24°C. Buffer in this assay was buffer B were passed through a propyl-agarose (Sigma)
(see under 'Enzyme purification' below). column (0.7 cm x 5cm) equilibrated with buffer B.
The protein peak that appeared was collected.
Measurement of ATP cleavage Step 4. Enzyme from the propyl-agarose column
[14C]ATP (0.5yCi) or 0.24uCi of [32P]ATP were was passed through a DEAE-cellulose column (DE
adjusted with unlabelled ATP to 1OpM in a total 52, Whatman; 1 cm x 4cm) equilibrated with buf-
volume of 50pl in the presence of enzyme, 0.5mM- fer B. Enzyme was not adsorbed on the ion-ex-
ergoline and 1 mM-MgCl2. Incubation was carried changer, and the protein peak that appeared was
out at 24°C for up to 30min. Portions of the collected.
D-Lysergic acid activation in Claviceps purpurea 859
Specific activity Protein determinations
Specific activities are expressed in nkat (1 nkat These were done as described by Bradford
is the amount of enzyme catalysing the incorpora- (1976) or Warburg & Christian (1941).
tion of 1 nmol of pyrophosphate into ATP/s.
M, determinations Radioactivity was determined with a Tri-Carb
liquid-scintillation counter (Packard Instruments).
The M, of the native enzyme was estimated by Charcoal filters were directly counted for radio-
gel filtration on a column (1.6cm x 100 cm) of activity in a toluene-based scintillation fluid.
Ultrogel AcA 34 using buffer B. The column was Radioactivity on t.l.c. plates was analysed by auto-
calibrated with gramicidin S synthetase II (Mr radiography using Kodak X-Omat S X-ray film.
280000), enniatin synthetase (250000), aldolase
(158000) and bovine serum albumin (68000) (see Strain and culture
Fig. 2). C. purpurea strain 1029 is a derivative of C.
purpurea A.T.C.C. 20102 (Keller, 1983). It was
maintained on slants of T 2 (Amici et al., 1967).
Cultivation in liquid medium was done as de-
scribed previously (Keller et al., 1980). For enzyme
Co preparation, 5-7-day-old cultures were used (300-
=CL 500mg of ergot peptide/litre). Freeze-dried myce-
-20 C. , lia from those cultures could be stored for at least
1 v0 3 months at -20°C.
X CZ Results
OI Enzyme purification
0-1_ By means of the D-lysergic acid-dependent
0 10 20 30 40 50 60 ATP-pyrophosphate exchange reaction it was
Fraction no. possible to purify a D-lysergic acid-activating
Fig. 2. Ultrogel AcA 34 gel filtration of crude enzyme enzyme about 145-fold. As is illustrated in Table 1,
extract from Table I an enzyme extract obtained by DEAE-cellulose
A 4ml portion of enzyme solution was fractionated fractionation of a crude protein extract was
on a column of dimensions 100cm x 1.6cm. Flow applied to an Ultrogel AcA 34 column (Fig. 2).
buffer was buffer B. Fractions of volume 3.3 ml were This separation resulted in a 47-fold increase of
collected. - -, A280; 0, D-lysergic acid-dependent specific activity accompanied by a 42-fold increase
pyrophosphate exchange; 0, D-lysergic acid-inde- in total activity. The latter effect was mainly due to
pendent pyrophosphate exchange. Arrows indicate
positions of various marker proteins: 1, gramicidin the removal of a non-specific ATP-PPi exchange.
S synthetase heavy enzyme (Mr 280000); 2, enniatin Thus the D-lysergic acid-dependent ATP-PP-
synthetase (250000); 3, aldolase (158000); 4, bovine exchange was almost completely free of back-
serum albumin (68000). ground activity (<5% of the total).
Table 1. Purification of D-lysergic acid-activating enzyme
Enzyme was purified from 6.5g of freeze-dried mycelium.
Total Total Specific
protein activity activity Yield
Purification step (mg) (nkat) (nkat/mg) (%)
1. C.rude enzyme extract 72 0.15 0.00208 100
pH 6, 5mM-sodium
2. Ultrogel AcA 34 64 6.3 0.0977 4200
3. Propyl-agarose 19.4 2.03 0.1047 1350
4. DEAE-cellulose, 6.7 2.03 0.3036 1350
860 U. Keller, R. Zocher, U. Krengel and H. Kleinkauf
The next steps of protein purification (see the methyl ester, dihydrolysergic acid amide, chano-
Materials and methods section) involved passages clavine, agroclavine, elymoclavine and ergotamine
of the pooled AcA 34 fractions through propyl- did not react under these conditions. This finding
agarose and DEAE-cellulose, by which a final was not surprising with respect to all of the ergo-
specific activity of 0.3nkat/mg of protein was lines lacking a carboxy group, but could not be
attained. The overall purification was 145-fold. At understood with respect to dihydrolysergic acid in
that point the protein fraction contained no view of the fact that this compound is incorporated
measurable background activity. For routine prep- into dihydroergotamine in vivo (Anderson et al.,
arations, step 3 was omitted because 60% of the 1979; U. Keller, unpublished work). As we argued
total activity was usually lost during this step by as- that the reaction rate of the dihydrolysergic acid-
yet-unknown reasons. Also, in this case, enzyme dependent pyrophosphate exchange was too low
did not contain any measurable background to be measurable under the above conditions, these
activity. were changed by omitting unlabelled pyrophos-
phate from the reaction mixtures. The time courses
Activation reaction of the reactions that were dependent on D-lysergic
The ATP-pyrophosphate exchange reaction acid and dihydrolysergic acid (as given in Fig. 3)
dependent on D-lysergic acid proceeded linearly clearly reveal that, under these conditions, a signi-
with time up to 30min in the presence of 0.25mM- ficant dihydrolysergic acid-dependent ATP-pyro-
pyrophosphate. There was also a strict proportion- phosphate exchange took place, which increased
ality between the reaction rate and the amount of with time up to 25 min. The enormous reactivity of
enzyme added. However, dihydrolysergic acid and D-lysergic acid compared with that of dihydro-
a variety of other ergolines such as D-lysergic acid lysergic acid is well illustrated by the observation
that nearly all of the radioactive pyrophosphate
present in the incubation mixture with D-lysergic
acid was incorporated into ATP during the first
2.5 min of the incubation (Fig. 3).
Further evidence for the ability of the enzyme to
activate both D-lysergic acid and dihydrolysergic
E 2( acid came from experiments in which chemically
synthesized adenylates of the two acids were
-0 incubated with [32P]pyrophosphate in the presence
or absence of enzyme. From Fig. 4 it is clear that
considerable ATP synthesis took place in the
presence of D-lysergic acid adenylate, whereas
C.) dihydrolysergic acid adenylate gave a small but
significant amount of ATP. No radioactive ATP
:^ was detected when D-lysergic acid or dihydro-
lysergic acid adenylates were replaced by the acids
X plus AMP or when the enzyme was omitted from
x those incubations containing adenylates (the latter
not shown). Attempts to isolate enzymically
formed adenylates from the incubation mixtures
by means of a variety of methods (Eigner &
Loftfield, 1974; Jakubowski et al., 1977) failed.
0 5 10 15 iO 25
Time (min) Purther enzyme properties
Fig. 3. Time course of ATP-pyrophosphate exchange Measurements of the influence of various D-
dependent on )-lysergic acid (O), dihydrolysergic acid (0) lysergic acid concentrations on the rate of the pyro-
and buffer (EQ) phosphate exchange reaction revealed an apparent
An 800pM portion of enzyme solution was incubated Km value between O.1 mm and 0.2mM under the
in a total volume of 1.5nml in the presence of conditions described. The enzyme's M, was deter-
1.55 x 106c.p.m. of [32P]pyrophosphate, 0.5 mM-D- mined by gel filtration on Ultrogel AcA 34 (Fig. 2).
lysergic or dihydrolysergic acid. All other conditions By the use of several standard proteins, its value
were as described in the Materials and methods
section, with the exception that unlabelled pyro- was estimated to lie between 135000 and 140000.
phosphate had been omitted. The control contained When the enzyme was incubated with [14C]- or
neither of the two acids. At the indicated times [32P]-ATP in the presence or absence of D'lysergic
200#1 portions were removed and added to I ml acid, no radioactive AMP nor ADP was formed.
portions of Norit A suspension. This indicates that no covalent bond between the
D-Lysergic acid activation in Claviceps purpurea 861
If D-lysergic acid is a free intermediate in
lysergyl-peptide formation, prior activation of the
carboxy group of this compound must take place.
The results presented here provide strong evidence
that this activation is accomplished via the
adenylate. Such evidence came from the finding
that the activating enzyme catalyses the formation
of [32P]ATP from [32P]pyrophosphate and chemi-
cally synthesized D-lysergic acid adenylate and
dihydrolysergic acid adenylate as well. In addition,
there is a strong specificity of the enzyme for its
substrate, D-lysergic acid, as revealed by its ability
to catalyse the ATP-pyrophosphate exchange
reaction. Together with the finding that dihydro-
lysergic is accepted to a slight, but significant,
Start- extent, these results substantiate the suggestion
that the enzyme plays an active role in the biosyn-
1 2 3 4 thesis of ergot peptides. Previous work revealed
Fig. 4. (3 2P]A TP synthesis from synthetic D-lysergic (or that D-lysergic acid and dihydrolysergic are effec-
dihydrolysergic) acid adenylates and [32PJpyrophosphate tively incorporated into certain derivatives in vivo
About lOO0pg of adenylates in each experiment were (Anderson et al., 1979; Keller et al., 1980;
used. The controls contained 5Onmol each of D- Willingdale et al., 1983), and it is likely that the
lysergic or dihydrolysergic acid and AMP instead of free acids serve as the precursors of peptide-bound
adenylates. Chromatography was carried out on ergolinic acids.
PEI-cellulose with solvent system IV. Co-chromato-
graphy of ATP revealed the identity of the Although the Mr of the enzyme (135000-
compounds. Exposure to X-ray ifim W,~as for 1 day. 1, 140000) seems to be high enough for harbouring
D-Lysergic acid adenylate; 2, dihydrolysergic acid more than one catalytic function (i.e. activation)
adenylate; 3, D-lysergic acid plus AMP; 4, dihydro- (see Lipmann, 1971), no evidence was obtained
lysergic acid plus AMP. that the enzyme is able to catalyse the formation of
a D-lysergyl-peptide or even a D-lysergyl-CoA thio-
ester. This latter compound has been claimed to be
a precursor of peptide-bound D-lysergic acid
(Maier et al., 1972). Thus it must be supposed that
another compound serves as an acceptor for the
enzyme and D-lysergic acid is formed, and another activated D-lysergic acid residue. Certainly the
molecule serves as acceptor of the D-lysergic acid enzyme does not form a covalent enzyme-sub-
strate complex, since no AMP release from ATP
residue. On the other hand, CoA or acetyl-CoA was observed, in either the presence or the absence
had no measurable influence on AMP release in of D-lysergic acid. It may be speculated that D-
the presence or absence of D-lysergic acid, and no lysergic acid adenylate reacts either with pre-
evidence was obtained for formation of a D- formed Ala-Phe-Pro tripeptide or with another
lysergyl-alanine or -valine peptide through enzy- protein, as in the case with D-phenylalanine in the
mic action (by the use of 14C-labelled amino acids). synthesis of the cyclic decapeptide gramicidin S
In addition, there was no indication of activation (Lipmann, 1971). Therefore further experiments
of any one of the amino acids present in the peptide will be required to elucidate the nature and
chains of ergot peptides produced by C. purpurea sequence of reactions between D-lysergic acid
A.T.C.C. 20102 or for activation of tryptophan. activation and D-lysergyl-peptide formation.
The enzyme functioned over a pH range between
6.5 and 9 with a slight optimum at pH 8 (measured We thank Dr. H. Kobel, Sandoz AG, Basel, Switzer-
in 100mM-Tris/HCl/3mM-dithioerythritol/0.5mM- land, for supplying us with various ergolines and Mr. M.
PMSF). At pH6.8 (buffer B), enzyme had 65% of Han for helpful discussions. This work was supported by
the activity of that measured at pH 8, but it was the 'Deutsche Forschungsgemeinschaft' (Sfb 9, D 5).
more stable at the lower pH (no loss of activity
after 2 days on ice compared with 30% reduction at
pH 8). Enzyme could be stored in 20% (v/v) References
glycerol at - 18°C for more than 6 months without Amici, A. M., Minghetti, A., Scotti, T., Spalla, C. &
significant loss of activity. Tognoli, L. (1967) Appl. Microbiol. 15, 597-602
862 U. Keller, R. Zocher, U. Krengel and H. Kleinkauf
Anderson, J. A., Kim, I. S., Lethonen, P. & Floss, H. G. Lee, S. G. & Lipmann, F. (1975) Methods Enzymol. 43,
(1979) J. Nat. Prod. 42, 271-273 585-602
Berg, P. (1958) J. Biol. Chem. 233, 608-611 Lipmann, F. (1971) Science 173, 875-884
Bradford, M. M. (1976) Anal. Biochem. 72, 248-254 Maier, W., Erge, D. & Gr6ger, D. (1972) Biochem.
Eigner, E. A. & Loftfield, R. B. (1974) Methods Enzymol. Physiol. Pflanz. 163, 432-442
29, 601-620 Moldave, K., Castelfranco, P. & Meister, A. (1959) J.
Groger, D. (1975) Planta Med. 28, 37-51 Biol. Chem. 234, 841-848
Hofmann, A. (1964) Die Mutterkornalkaloide, p. 116,
Verlag Enke, Stuttgart Ru6man, R. (1976) J. Chromatogr. 121, 353-360
Jakubowski, H. Z., Pastuzyn, A. & Loftfield, R. B. Stoll, A., Hofmann, A. & Becker, B. (1943) Helv. Chim.
(1977) Anal. Biochem. 82, 29-37 Acta 26, 1602-1613
Keller, U. (1983) App!. Environ. Microbiol. 46, 580-584 Warburg, 0. & Christian, W. (1941) Biochem. Z. 310,
Keller, U., Zocher, R. & Kleinkauf, H. (1980) J. Gen. 384-421
Microbiol. 118, 485-495 Willingdale, I., Atwell, S. M. & Mantle, P. G. (1983) J.
Kleinkauf, H. (1979) Planta Med. 35, 1-18 Gen. Microbiol. 129, 2109-2115