Eur. J. Biochem. 196,687-692 (1991) 0 FEBS 1991 0014295691001933
Enzyme-mediated peptide synthesis using acylpeptide hydrolase Timothy C. FARRIES', Anthony D. AUFFRET' and Alastair AITKEN3
'
'
MRC Collaborative Centre, London, England Pharmacia LKB Biochrom Ltd, Cambridge, England The National Institute for Medical Research, London, England
(Received September 19/December 10, 1990) - EJB 90 1123
Acylpeptide hydrolase is shown to catalyse the specific addition of a single amino acid to the N-terminus of a peptide. The stabilised Sepharose-coupled form of the enzyme is used to couple a carboxy-methylated N-formyl (or N-acetyl) amino acid to a short pre-existing peptide. The yield is improved by optimal timing of the reaction and the presence of moderate concentrations (5%) of N,N-dimethylformamide. Two tripeptides, Ac-Ala-Ala-Ala and Wet-Leu-Phe (f, formyl) were synthesized by this technique (in yields of 2% and 0.064% respectively). The products were characterised by HPLC, amino acid analysis, mass spectroscopy and protein sequencing. The synthetic Met-Leu-Phe also had biological activity, in that it stimulated superoxide generation by granulocytes. Acylpeptide hydrolase could therefore be a very useful tool for the synthesis and modification of peptides.
Most peptides are currently synthesized entirely by chemical methods using a strategy based either on the procedure of Merrifield [ 11 or the 9-fluorenylmethoxycarbonyl-labelled amino acid method devised by Atherton and Sheppard [2]. Amino acids are sequentially coupled to the N-terminus of a chain attached to a solid support by its C-terminus. Advantages of this method include flexibility, high yields and adaptability to automation. However, chemical peptide synthesis also has disadvantages, particularly the possibilities of racemization and serial side reactions. Elaborate chemistry is required to protect all the reactive side chains making the method unsuitable for extending pre-existing peptides. In contrast enzyme-catalysed peptide synthesis is intrinsically regiospecific and does not require chemical protection of amino acid side chains (reviewed in [3,4]). Proteases, which normally hydrolyse peptide bonds, are used to catalyse this reaction in reverse. Reactions are driven towards synthesis using activated carboxyl derivatives, organic cosolvents and kinetic or physicochemical trapping of the product. Endopeptidases such as trypsin and chymotrypsin have been used, but their applicability is severely restricted by their specificity. Far more generally useful would be enzymes that could add amino acids sequentially to one terminus of a peptide. However, exopeptidases can catalyse not only the addition of one residue, but also the addition of more than one residue in the same step, or the removal of terminal residues from the Correspondence to A. Aitken, MRC Collaborative Centre, 1 - 3 Burtonhole Lane, Mill Hill, London, NW7 IAD, England Abbreviations. APH, acylpeptide hydrolase (also known as acylamino-acid-releasing enzyme); DAPH-S, APH coupled to Sepharose and treated with dimethylsuberimidate; M e t , formylmethionine; OMe, methyl ester; FAB-MS, fast-atom-bombardment mass spectroscopy. Enzymes. Acylpeptide hydrolase (EC 3.4.19.1); superoxide dismutase (EC 1.15.1.1). Noze. A patent application concerning the use of the technology described in this paper has been sent to the Swedish patent office [20].
substrate peptide. In some cases it has been possible to extend the C-terminus of a peptide by using a carboxypeptidase and an amino acid with a blocked carboxyl group. Aminopeptidases require the amino groups of both the substrate peptide and the incoming amino acid to be unmodified and serial additions are possible. The reaction of residue A with a peptide BCD would therefore yield a mixture of products including ABCD, AABCD, CD and ACD. Acylpeptide hydrolase (APH, also known as acylaminoacid-releasing enzyme) is an exopeptidase that removes Nacetylated or N-formylated N-terminal residues from peptides, but does not degrade peptides with free N-termini [591. In this paper, we have used the properties of this enzyme to overcome the disadvantages described above for other proteases. Firstly the substrate peptide has a free N-terminus and so is not a substrate for cleavage by APH. Secondly, the product peptide is N-terminally acylated, preventing the unwanted addition of multiple residues in any one step. The scheme used is shown in Fig. 1. A carboxyl-esterified acylamino acid will transesterify on to the enzyme, then react with the substrate peptide (or with HzO). Methyl esters were used as they are known to be good substrates for APH [5]. Formyl amino acids have the advantage over the acetylated derivatives that the formylated products can be unblocked in a separate step by mild acid hydrolysis. In this paper we describe the initial development of the method and demonstrate its feasibility by synthesising tripeptides and dipeptides. The products were identified by physical and chemical characterization. The enzymatically synthesized Met-Leu-Phe was also shown to have biological activity. MATERIALS AND METHODS Reagents
CNBr-activated Sepharose 4B was a product of Pharmacia-LKB (Uppsala, Sweden). Dimethylsuberimidate
688 X-NH-R,-OMe
+ I
Amino acid and protein sequence analysis
APH
I
t X-NH-R,-OH
(X = formyl or acetyl )
f
X-NH-R mRn+,----OH
I
H
+
(X = formyl)
NH*-R,-Rn+r ----OH
Fig. 1. Strategy,for peptide synthesis using acylpeptide hydrolase. The figure shows the general scheme for coupling a residue R, onto the existing N-terminus of a peptide. Deformylation of the product is performed as a separate step by mild acid hydrolysis
was purchased from Pierce (Rockford, IL, USA), acetic anhydride, benzene and trifluoracetic acid from BDH Ltd (Poole, UK) and HPLC-grade reagents (acetonitrile, acetic acid and methanol) from Romil Chemicals Ltd (Loughborough, UK). Products of Sigma Chemical Co. (St Louis, MO, USA) included N,N-dimethylformamide, Histopaque 1077, Histopaque 1119 and the following peptides and amino acid derivatives : fMet-Leu-Phe, (f, formyl), AcAAA (Ac-Ala-AlaAla, Ala-Ala, Ac-Ala, Leu-Phe, fLeu, Met-OMe (methylmethioninate), Phe-OMe. ~-[U-l~C]Alanine (6 GBq/mmol) was purchased from Amersham International plc (Amersham, UK). Preparation qf madijied amino acids and peptides N-Acetylation of amino acids and peptides was performed in 1 :1 mixtures of acetic anhydride and acetic acid at 22 "C for 1 h, followed by evaporation in a centrifugal evaporator or by lyophilisation [lo, 111. N-Formylation of amino acids was similarly performed in 3 : 1 mixtures of formic acid/acetic anhydride at 22 C for 1 h [12, 131. Methyl esters of amino acids were prepared in 0.1 M methanol/HCl, at 22°C for 16 h and the products dried in vucuo [14]. Preparation q f A P H and A P H linked to Sepharose and treated with ~liw?ethylsubrrimidate( D A P H - S ) APH was purified from ovine liver as previously described [6, 151. The enzyme was then coupled to CNBr-activated Sepharose and cross-linked with dimethylsuberimidate as described elsewhere [15]. This preparation is known as DAPH-S. Reverse-phase HPLC All HPLC separations were performed on the System Gold (Beckman, Palo Alto. CA, USA) with a CI8 reverse-phase column (220 mm = 2.1 mm, 5 pm, supplied by Applied Biosystems, Warrington. UK). The flow rate was 0.3 ml/min and the mobile phase was 0.1% aqueous trifluoroacetic acid. Gradients of CHJCN were used as indicated.
For amino acid analysis samples were hydrolysed with 6 M HC1, in the vapour phase, at 110"C for 18 - 20 h, before analysis by an Applied Biosystems 420A amino acid analyzer with on-line model 130A PTC analyser. Protein sequences were determined by automated Edman degradation in an Applied Biosystems 477A protein sequencer with an on-line phenylthiohydantoin amino acid analyser. Fast-atom-bombardment mass spectroscopy ( F A B - M S ) FAB-MS was performed in the positive-ion mode in a VG 250-70 SE instrument, using a 30-keV caesium ion gun. Thioglycerol plus 1% trifluoroacetic acid was used as matrix material. Stimulation of superoxide generation by granulocytes Human granulocytes were purified from 30 ml fresh peripheral blood by overlaying the diluted blood on to layered cushions of Histopaque 1077 and Histopaque 1119 followed by centrifugation at 700 x g for 30 min (according to the manufacturer's instructions). The cells were taken from the interface, washed and used at lo6 viable cells/l ml sample (by trypan blue exclusion) in Krebs-Ringer phosphate solution. Stimulation of superoxide release was measured from the superoxide-dismutase-inhibitable reduction of cytochrome c, as described by Greenberger et al. [16]. Each reaction was performed for 20 min at 37°C in the presence of 0.05 mM cytochrome c and the presence or absence of 0.01 mg/ml superoxide dismutase. The absorbances of the supernatants were meaured at 550 nm. The background absorbances from (a) corresponding superoxide dismutase containing samples, and (b) the negative control with no peptide were subtracted. Hence superoxide production was calculated by dividing by the molar extinction coefficient for the reduction of cytochrome c (21 000 M-'). RESULTS Optimization of conditions f o r synthesis of Ac-Ala-Ala-Ala Ac-Ala-Ala-Ala is known to be a very good substrate for acylpeptide hydrolase [7 - 91. The conditions required for peptide synthesis were therefore initially characterized for the following reaction : Ac-[14C]Ala-OMe MeOH.
+
Ala-Ala
Ac-[14C]Ala-Ala-Ala
+
Use of a large excess of nonradioactive Ala-Ala allowed the radioactive peptide product to be detected with high sensitivity. A time-course of the reaction at 22 "C in aqueous solution is shown in Fig. 2. Hydrolysis to A ~ - [ ' ~ c ] A lwas a the predominant reaction at all time points. Synthesis of Ac['4C]Ala-Ala-Ala reached a plateau at 30 - 60 min, after which synthesis is presumably exceeded by secondary enzymic hydrolysis to Ac-[14C]Ala + Ala-Ala. At no point did the yield exceed 5% (of radioactive component), and at 30 min the amount of Ac-[14C]Ala-Ala-Ala generated was only 16% of the yield of the unwanted product, Ac-[14C]Ala. In order to improve the yield, the effects of organic cosolvents were studied. We have already shown that APH retains some activity in low concentrations (5%, by vol.) of N, N-dimethylformamide, and that N,N-dimethylformamide
689
f 11 .-
m
c c
I
0
I
5
I
15
I
30
Time (min)
I 60
are increased accordingly. Further studies were therefore performed in the presence of unlabelled Ac-Ala-OMe (1 mM). These demonstrated that enzyme action could be accelerated at 45 "C with a corresponding improvement in yield, and that the ratio of A~-['~C]Ala-Ala-Ala/Ac-[~~C]Ala generated was further increased by emulsifying the solution (in the presence of a nonionic detergent) with an organic phase of benzene (not shown). In the latter case, it was subsequently found that Ac-Ala-Ala-Ala did not significantly partition into the benzene phase, so the effect was probably due to the partial solubility of benzene in the aqueous phase, acting like N,Ndimethylformamide to decrease the polarity and H 2 0 activity of the reaction medium. Production of A~-[~~C]Ala-Ala-Ala could thereby be increased to 32% of Ac-[14C]Ala. In this respect, octane (less polar than benzene) had less effect whereas butanol (more soluble in water) inhibited the enzyme.
Fig. 2. Time course of the production of Ac-['4C]Ala-Ala-Ala and Ac[14C]Ala. The reaction mixture contained 1 nmol A c - [ ' ~ C ] A I ~ - O M ~ (3 x lo5 cpm), 50 pmol Ala-Ala and 2 p1 DAPH-S (60 ng APH) in 0.3 ml50 mM NaH2P04/Na2HP04,pH 8, buffer and was incubated Enzymatic synthesis of Ac-Ala-Alu-Ala at 22°C. At the times indicated, 0.05 ml samples were taken, mixed with 0.4 ml ice-cold 0.1% trifluoroacetic acid and briefly centrifuged A larger scale synthesis of Ac-Ala-Ala-Ala was attempted (10 s, 16000 xg). The supernatants were concentrated by freeze-dry- using conditions derived from the preceding results. The reacing then analysed by reverse-phase HPLC in aqueous 0.1% tion mixture (RI) contained the following components : Actrifluoroacetic acid (isocratic conditions) as described in Materials Ala-OMe, 0.5 pmol; Ala-Ala, 50 pmol ; N,N-dimethyland Methods. Fractions corresponding to A ~ - [ ' ~ c ] A l(5 a -6 min, formamide, 5% (by vol.); Nonidet P40, 1% (by vol.); W ) and A~-['~C]Ala-Ala-Ala(11 -13 min, 0 ) were collected and DAPH-S (enzyme), 10 p1 (0.3 pg APH). All these were disradioactivity measured
solved in 0.25 ml 50 mM NaH2P04/Na2HP04,pH 8, mixed with 0.5 ml benzene. A negative control (RO) was included 20 with identical conditions except for the absence of enzyme (DAPH-S). After shaking for 2 h at 45"C, DAPH-S was removed by centrifugation and the supernatant added to 0.5 ml 0.1% (by vol.) trifluoroacetic acid. This inactivated any residual enzyme. Both samples were dried in vucuo and the 15products separated by reverse-phase HPLC (as described in Materials and Methods). The fractions corresponding to the a 10elution position (11- 13 min) of Ac-Ala-Ala-Ala were col8 lected, freeze-dried in four equal aliquots and analysed as a a follows. Q, 5HPLC. One aliquot of each sample was dissolved in 0.1 % a trifluoroacetic acid and refractionated by HPLC as above. R1 contained a peak that co-eluted with an Ac-Ala-Ala-Ala standard and was absent in RO (not shown). This peak had I I I I I been partially obscured by overlap with the much larger peak 0 5 10 20 30 Dimethyl Formamide (%) of Ala-Ala in the original HPLC fractionation of the products, Fig. 3. Effect of",N-dimethylformamide on synthesis of A c - [ ' ~ C ] A I ~ - but was totally resolved in the second chromatography. Alu-Alu. The reaction mixture contained 0.1 nmol A c - [ ' ~ C ] A I ~ - O M ~ FAB-MS. A portion of each sample was analysed by FABMS, as shown in Fig. 4.The predicted ion mass of 274, [Ac(60000 cpm), 10 pmol Ala-Ala, 2 pl DAPH-S (60 ng APH) and the Ala-Ala-Ala HIf in the positive-ion mode, was evident in indicated concentrations of N,N-dimethylformamide in 0.1 ml50 mM NaH2P0,/Na2HP04, pH 8, buffer. After 60 min at 2 2 T , 0.5 ml the sample incubated with enzyme but negligible in the nega0.1% trifluoroacetic acid was added to each. The samples were then tive control. This confirmed that the correct product had been centrifuged, freeze-dried and analysed by HPLC as described in Fig. 2. produced. The generation of A~-['~C]Ala-Ala-Ala is shown as a product ratio Amino acid analysis. One of the aliquots of each sample (a percentage (0)and as total percentage yield ( 0 ) quarter of the total) was subjected to amino acid analysis as described in Materials and Methods. The results indicated 10.33 nmol Ala in R1 compared to 2.98 nmol in RO (alanine is present in RO after HPLC due to overlap with the peak resistance is enhanced in the modified enzyme, DAPH-S [15]. of Ala-Ala). No other residues were present in significant The effect of N,N-dimethylformamide on Ac-[14C]Ala-Alaamounts. The total yield of Ac-Ala-Ala-Ala was therefore Ala synthesis by DAPH-S is shown in Fig. 3. The yield was calculated as 4 x (10.33 - 2.98)/3 = 9.8 nmol. This represents improved in 5% N,N-dimethylformamide, but reduced in a yield of 2% of the Ac-Ala-OMe substrate. higher concentrations, presumably due to inactivation of the enzyme. More importantly for the objective of scaling up the process, 5% N,N-dimethylformamide improved the product Enzymatic synthesis o jfMet-Leu-Phe ratio of Ac-['4C]Ala-Ala-Ala/Ac-[14C]Ala by almost fourfold. This peptide was selected for enzymatic synthesis for the The results predict that significant amounts of product should be obtainable if the amounts of substrate and enzyme following reasons : the formyl group can be removed by mild
d
5-
v
,J
+
acid hydrolysis allowing sequence analysis or addition of further residues; M e t is a favourable terminal group for APH action [6, 91; Met-Leu-Phe has a biological activity, being a potent example of the bacterial formylpeptides that stimulate chemotaxis and trigger a respiratory burst in mammalian granulocytes [17]. The reaction mixture (MLF1) contained the following components: met-OMe, 12.5 pmol; Leu-Phe, 17 pmol; N , N dimethylformamide, 5% (by vol.); Nonidet P40,8% (by vol.); DAPH-S (enzyme), 10 p1 (0.3 pg APH). All were dissolved in 0.25 ml 50 mM NaH2P04/Na2HP04, pH 8, mixed with 0.5 ml benzene. The reaction conditions were as described above for Ac-Ala-Ala-Ala, including the incorporation of a negative control without DAPH-S (MLFO). The products were fractionated by reverse-phase HPLC, as before, but with a 40-min gradient of 0 - 50% CH,CN (started after 10 min in aqueous 0.1 YOtrifluoroacetic acid). The resulting chromatograms are shown in Fig. 5. Only one enzyme-dependent peak is evident (labelled Product) and coincides with the elution volume of standard met-Leu-Phe. This fraction was collected from both samples, freeze-dried in four equal aliquots, and analysed as follows. FAB-MS. FAB-MS, shown in Fig. 6, identified an ion mass of 438 as predicted for [Wet-Leu-Phe HI+ in the positive-ion mode. This species was absent in the negative control. Amino acid analysis. One aliquot of each was subjected to amino acid analysis, yielding the results shown in Table 1. These clearly show that the product contains approximately 2 nmol each of Met, Leu and Phe, indicating a total yield of 8 nmol. This is a 0.064% yield from the substrate Ac-MetOMe. Conversion to Met-Leu-Phe. The formyl group was hydrolysed from one aliquot of each sample in 0.5 M methanol/HCl (24 h at 22°C). A portion of commercial met-Leu-Phe was identically treated in parallel. After drying in vacuo these samples were de-esterified with 0.25 M NaOH (22"C, 1 h) followed by neutralisation with 0.25 M HCl. The products were analysed by reverse-phase HPLC as above. The chromatograms, shown in Fig. 7, indicate that the product now eluted at 44 min (compared to 48 min before hydrolysis, see Fig. 5 ) , with no comparable product in the negative control. This peak was collected and used in sequence analysis as
+APH
100-
274 = IAcAAA+Hl+
1
50-
I
-APH
+
0-I 100
L
200
250
300
400
350
MIZ
Fig. 4. Mass spectrurn of the product of Ac-Ah-Ala-Ah synthesis. FAB-MS in the positive-ion mode was performed on the purified products of the synthesis of Ac-Ala-Ala-Ala under the conditions given in Materials and Methods. The mass ion at 274 corresponds to the m / z (M/Z) predicted for [Ac-Ala-Ala-Ala + HIi, and the ion at 296 corresponds to the sodium adduct [Ac-Ala-Ala-Ala + Na] . The spectrum of the negative control sample without enzyme (-APH) is also shown. The peak at mjz 381 is derived from the thioglycerol matrix and is present in both samples +
I
I 20
I
I
I
I
I
I
I
I
I 60
Time (rnin)
Fig. 5. Sepurution of the products of fMet-Leu-Phe synthesis by reverse-phase HPLC. The reaction conditions for the synthesis of met-LeuPhe are reported in the text. The HPLC conditions are as described in Materials and Methods using a 40-min gradient to 50% (by vol.) CH,CN/O.l% trifluoroacetic acid started after running for 10 min in 0.1 % trifluoroacetic acid. The chromatogram of the negative control sample (-APH) is also shown. The eluate at the position labelled Product (48 min) was collected from both separations for further analysis
69 1 438
=
[fMLF
+
HI+
"'1
+APH (MLF1)
I
a
I
I
I
I
I
I
I
I
I
I B
-0.022
C
1'291 For-MLF
h i l l I
1 -APH (MLFO)
nn Y."
-
I
0
lo] 50
350
450
400
500
5 50
I
I
I
I
Time (rnin)
I
I
I
I
I
60
Fig. I . HPLC analysis of the product of fMet-Leu-Phe synthesis after mild acid hydrolysis. The samples from fMet-Leu-Phe synthesis (+ APH; A), the negative control (- APH; B) and commercial M e t Leu-Phe (For-MLF; C) were prepared and hydrolysed in methanol/ HCI as described in the text. These were then analysed by reversephase HPLC using the same conditions described in Fig. 5 . Successful conversion of fMet-Leu-Phe to Met-Leu-Phe is indicated by the shift in elution position from 48 rnin (in Fig. 5 ) to 44 rnin
I
300
I
600
MI2
Fig. 6. Mass spectrum of the product offMet-Leu-Phe synthesis. Positive-ion FAB-MS was performed on the purified products of the synthesis of fMet-Leu-Phe (described in the text) under the conditions given in Materials and Methods. The spectrum of the negative control sample (MLFO) is also shown. The ion mass of 438 corresponds to the m / z (M/Z) predicted for [Wet-Len-Phe + HI+. The other prominent signals are present in both spectra
bined with the mass spectral data, this confirms that the correct product has been produced. Activity assay. Biological activity of the products was determined from their ability to stimulate the respiratory burst in human granulocytes (as described in Materials and Methods). The results, shown in Table 3, show that the product has a total biological activity slightly greater than the equivalent amount of commercial fMet-Leu-Phe.
Other synthetic products Table 1. Amino acid analysis of the product of fMet-Leu-Phe synthesis Amino acid analysis was performed on an aliquot of each sample as described in Materials and Methods. All other residues detected were present in less than 0.1 nmol. Cys and Trp were not determined _
_
~ ~~~
Amino acid
Amino acids in fMet-Leu-Phe MLFI
MLFO
Leu Phe Met
2.2 2.1 1.5
0.1 0.1 0.0
Ala Are GIY
0.3 0.2 0.1
0.4 0.0 0.1
described below. The standard fMet-Leu-Phe showed a similar shift in elution position after hydrolysis. Sequence analysis. The HPLC-purified deformylated product was freeze-dried and sequenced by automated Edman degradation. The result, shown in Table 2, gives an unambiguous confirmation that the sequence is Met-Leu-Phe. Com-
The range of potential applications was further illustrated by synthesis of the following two dipeptides. The products were identified by HPLC, mass spectroscopy and amino acid analysis. fMet-Met-OMe. The reaction mixture contained the following components: fMet-OMe, 1 pmol; Met-OMe, 10 pmol; N,N-dimethylformamide, 10% (by vol.); DAPH-S, 50 pl. All were dissolved in 0.25 ml50 mM NaH2P04/Na2HP04,pH 8 (without a benzene phase). After 1 h at 45"C, the supernatant was collected, dried in vucuo and fractionated by reverse-phase HPLC (using the conditions described above for Met-LeuPhe). A peak eluting at 40 rnin that was absent in the negative control (without DAPH-S) was collected, dried and analyzed by mass spectroscopy. An ion mass of 323, corresponding to [fMet-Met-OMe + HI', was observed and was not present in the corresponding fraction from the negative control. fLeu-Phe-0 Me. The reaction mixture contained the following components: fLeu-OMe, 12.5 pmol; Phe-OMe, 12.5 pmol; N,N-dimethylformamide, 5% (by vol.); DAPH-S, 50 pl. All were dissolved in 0.25 ml 50 mM NaH2P04/ Na2HP04, pH 8. After 2 h at 45"C, the supernatant was collected, dried in vucuo and fractionated by reverse-phase HPLC (using the conditions described above for Met-Leu-Phe). A peak eluting at 45 min that was absent in the negative control
Table 2. Amino acid .sequence of the product ofjMet-Leu-Phe synthesis Protein sequence analysis was performed on the products after acid hydrolysis and HPLC purification as described in the text and in Materials and Methods. A single unambiguous sequence was obtained (Met-Leu-Phe) Cycle
Amino acid
Amount pmol
1 2 3
Met Leu Phe
30 35 20
Table 3. Biologicul uctivitj of the product of fMet-Leu-Phe synthesis The sample were assayed for biological activity, as described in Materials and Methods by measuring their ability to stimulate the superoxide generation by human granulocytes. The results shown are the means from triplicates (fSEM), compared with the activity of a standard dilution of commercial Met-Leu-Phe Sample
Amount added
Superoxide generation pmo1/lO6 granulocytes
MLFI (enzymatic synthesis) MLFO (control) fMLF (commercial)
1.25% total 1.25% total 0.1I nmol
4.06 f 0.09 0.83 0.23 3.07
(without DAPH-S) was collected and dried. One half was subjected to amino acid analysis, yielding 75 pmol leucine and 117 pmol phenylalanine that were not present in the equivalent fraction of the negative control. The total yield of Ceu-Phe was therefore approximately 0.1 nmol.
DISCUSSION The application of acylpeptide hydrolase in peptide synthesis has been demonstrated by the detailed characterization of the production of two tripeptides (Ac-Ala-Ala-Ala and Met-Leu-Phe). The synthesis of two dipeptides (met-MetOMe and fLeu-Phe-OMe) is also reported. APH can thus be used to elongate peptide chains by sequential addition to the N-terminus. The enzymatically synthesised products are expected to contain only L-amino acids, although this has not been verified experimentally. Importantly, formylated products can be deprotected by mild acid hydrolysis to allow further elongation of the chain. As yet the yields have been too low to perform more than one cycle of amino acid coupling. There are more potential applications of the method than total peptide synthesis. APH could be used to modify preexisting peptides by, for example, introducing a reactive sidechain (e.g. cysteine) or a radiolabel. APH can also be used to remove acetylated or formylated N-terminal residues [5- 9, 151. These amino acids could then be mutated by using the same enzyme to remove residues and, under synthetic conditions, to add back others. It is also possible that APH could couple certain artificial amino acids ( e g homoserine and norvaline) which are difficult to introduce into proteins by genetic means.
The method does have a number of limitations. Firstly, the yields obtained have been poor. This could be improved simply by using much more enzyme but further development is also necessary. Secondly, although the enzyme has quite a broad substrate specificity, it is known not to recognise charged residues at the PI position [8, 9, 181. In total peptide synthesis APH may therefore have to be complemented by other enzymic or chemical methods. Finally the activity of APH declines sharply with increasing size of the substrate [9]. Perhaps this could be overcome by disrupting the secondary structure of the substrate, for example with detergents. The stability of the enzyme would have to be increased which would also allow use of higher concentrations of organic cosolvents. This would improve the yield by reducing the formation of the unwanted hydrolysis product. If the water content could be reduced to a minute fraction it would reduce the ionization of normally charged residues which might then become compatible with enzyme activity. A more stable enzyme might be obtained by chemically modification, genetic engineering or isolation from a thermophilic microorganism [19]. Acylpeptide hydrolase could then become a very useful tool in peptide chemistry. We are grateful to Alan Harris (NIMR, London) for the protein sequence analysis and to Fatima Beg (NIMR) for performing the FAB-MS and amino acid analysis. The financial support of Pharmacia LKB is also gratefully acknowledged.
REFERENCES 1. Merrifield, R. B. (1963) 1. Am. Chem. Soc. 85,2149-2154. 2. Atherton, E., Fox, H., Harkiss, D., Logan, C. J.. Sheppard, R. C. & Williams, B. J. (1978) J . Chem. Soc. Chem. Commun. 537 - 540. 3. Kullmann, W. (1985) J . Protein Chem. 4, 1-22. 4. Jakubke, H.-D. (1987) in Thepeptides. Analysis, synthesis, biology (Udenfriend, S . & Meienhofer, J., eds), vol. 9, pp. 103-165, Academic Press, San Diego. 5. Tsunasawa, S . , Narita, K. & Ogata, K. (1975) J . Biochem. (Tokyo) 77,89-102. 6. Tsunasawa, S . & Narita, K. (1976) Methods Enzymol. 45, 552561. 7. Gade, W. &Brown, J. L. (1978) J . Bid. Chem. 253, 5012-5018. 8. Jones, W. M. & Manning, J. M. (1985) Biochem. Biophys. Res. Cornmun. 126,933-940. 9. Kobayashi, K. & Smith, J . A. (1987) J . Riol. Chem. 262,1143511445. 10. Green, R. W., Ang, K. P. & Lam, L. C. (1953) Biochem. J . 54, 181 - 187. 11. Riordan, J. F. & Vallee, B. L. (1972) Methods Enzymol. 25,494499. 12. Sheehan, J. C. & Yang, D.-D. H. (1957) J . Am. Chem. Soc. 80, 1154- 1158. 13. Geiger, R. & Konig, W. (1981) in Thepeptides. Analysis, sjwthesis, biology (Gross, E. & Meienhofer, J., eds) vol. 3, pp. 1-99, Academic Press, New York. 14. Wilcox, P. E. (1967) Methods Enzymol. 11, 605-617. 15. Farries, T. C., Harris, A,, Auffret, A. D. & Aitken, A . (1991) Eur. J. Biochem. 196, 679-685. 16. Greenberger, J. S . , Newburger, P. E., Karpas, A. & Moloney, W. C. (1978) Cancer Res. 38,3340-3348. 17. Showell, H. J., Freer, R. J., Zigmond, S . H., Schiffmann, E., Aswanikumar, S . , Corcoran, B. & Becker, E. L. (1976) J . Exp. Med. 143,1154-1169. 18. Orfeo, T. & Meyer, W. L. (1986) Fed. Proc. 45, 1712. 19. Mozhaev, V. V., Berezin, I. V. & Martinek, K. (1988) CRC Crit. Rev. Biochem. 23,235-281. 20. Farries, T. C., Auffret, A. D. & Aitken, A. (1990) Swedish patent application No. SE 9000 597-6.
Enzyme-mediated peptide synthesis using acylpeptide hydrolase Timothy C. FARRIES', Anthony D. AUFFRET' and Alastair AITKEN3
'
'
MRC Collaborative Centre, London, England Pharmacia LKB Biochrom Ltd, Cambridge, England The National Institute for Medical Research, London, England
(Received September 19/December 10, 1990) - EJB 90 1123
Acylpeptide hydrolase is shown to catalyse the specific addition of a single amino acid to the N-terminus of a peptide. The stabilised Sepharose-coupled form of the enzyme is used to couple a carboxy-methylated N-formyl (or N-acetyl) amino acid to a short pre-existing peptide. The yield is improved by optimal timing of the reaction and the presence of moderate concentrations (5%) of N,N-dimethylformamide. Two tripeptides, Ac-Ala-Ala-Ala and Wet-Leu-Phe (f, formyl) were synthesized by this technique (in yields of 2% and 0.064% respectively). The products were characterised by HPLC, amino acid analysis, mass spectroscopy and protein sequencing. The synthetic Met-Leu-Phe also had biological activity, in that it stimulated superoxide generation by granulocytes. Acylpeptide hydrolase could therefore be a very useful tool for the synthesis and modification of peptides.
Most peptides are currently synthesized entirely by chemical methods using a strategy based either on the procedure of Merrifield [ 11 or the 9-fluorenylmethoxycarbonyl-labelled amino acid method devised by Atherton and Sheppard [2]. Amino acids are sequentially coupled to the N-terminus of a chain attached to a solid support by its C-terminus. Advantages of this method include flexibility, high yields and adaptability to automation. However, chemical peptide synthesis also has disadvantages, particularly the possibilities of racemization and serial side reactions. Elaborate chemistry is required to protect all the reactive side chains making the method unsuitable for extending pre-existing peptides. In contrast enzyme-catalysed peptide synthesis is intrinsically regiospecific and does not require chemical protection of amino acid side chains (reviewed in [3,4]). Proteases, which normally hydrolyse peptide bonds, are used to catalyse this reaction in reverse. Reactions are driven towards synthesis using activated carboxyl derivatives, organic cosolvents and kinetic or physicochemical trapping of the product. Endopeptidases such as trypsin and chymotrypsin have been used, but their applicability is severely restricted by their specificity. Far more generally useful would be enzymes that could add amino acids sequentially to one terminus of a peptide. However, exopeptidases can catalyse not only the addition of one residue, but also the addition of more than one residue in the same step, or the removal of terminal residues from the Correspondence to A. Aitken, MRC Collaborative Centre, 1 - 3 Burtonhole Lane, Mill Hill, London, NW7 IAD, England Abbreviations. APH, acylpeptide hydrolase (also known as acylamino-acid-releasing enzyme); DAPH-S, APH coupled to Sepharose and treated with dimethylsuberimidate; M e t , formylmethionine; OMe, methyl ester; FAB-MS, fast-atom-bombardment mass spectroscopy. Enzymes. Acylpeptide hydrolase (EC 3.4.19.1); superoxide dismutase (EC 1.15.1.1). Noze. A patent application concerning the use of the technology described in this paper has been sent to the Swedish patent office [20].
substrate peptide. In some cases it has been possible to extend the C-terminus of a peptide by using a carboxypeptidase and an amino acid with a blocked carboxyl group. Aminopeptidases require the amino groups of both the substrate peptide and the incoming amino acid to be unmodified and serial additions are possible. The reaction of residue A with a peptide BCD would therefore yield a mixture of products including ABCD, AABCD, CD and ACD. Acylpeptide hydrolase (APH, also known as acylaminoacid-releasing enzyme) is an exopeptidase that removes Nacetylated or N-formylated N-terminal residues from peptides, but does not degrade peptides with free N-termini [591. In this paper, we have used the properties of this enzyme to overcome the disadvantages described above for other proteases. Firstly the substrate peptide has a free N-terminus and so is not a substrate for cleavage by APH. Secondly, the product peptide is N-terminally acylated, preventing the unwanted addition of multiple residues in any one step. The scheme used is shown in Fig. 1. A carboxyl-esterified acylamino acid will transesterify on to the enzyme, then react with the substrate peptide (or with HzO). Methyl esters were used as they are known to be good substrates for APH [5]. Formyl amino acids have the advantage over the acetylated derivatives that the formylated products can be unblocked in a separate step by mild acid hydrolysis. In this paper we describe the initial development of the method and demonstrate its feasibility by synthesising tripeptides and dipeptides. The products were identified by physical and chemical characterization. The enzymatically synthesized Met-Leu-Phe was also shown to have biological activity. MATERIALS AND METHODS Reagents
CNBr-activated Sepharose 4B was a product of Pharmacia-LKB (Uppsala, Sweden). Dimethylsuberimidate
688 X-NH-R,-OMe
+ I
Amino acid and protein sequence analysis
APH
I
t X-NH-R,-OH
(X = formyl or acetyl )
f
X-NH-R mRn+,----OH
I
H
+
(X = formyl)
NH*-R,-Rn+r ----OH
Fig. 1. Strategy,for peptide synthesis using acylpeptide hydrolase. The figure shows the general scheme for coupling a residue R, onto the existing N-terminus of a peptide. Deformylation of the product is performed as a separate step by mild acid hydrolysis
was purchased from Pierce (Rockford, IL, USA), acetic anhydride, benzene and trifluoracetic acid from BDH Ltd (Poole, UK) and HPLC-grade reagents (acetonitrile, acetic acid and methanol) from Romil Chemicals Ltd (Loughborough, UK). Products of Sigma Chemical Co. (St Louis, MO, USA) included N,N-dimethylformamide, Histopaque 1077, Histopaque 1119 and the following peptides and amino acid derivatives : fMet-Leu-Phe, (f, formyl), AcAAA (Ac-Ala-AlaAla, Ala-Ala, Ac-Ala, Leu-Phe, fLeu, Met-OMe (methylmethioninate), Phe-OMe. ~-[U-l~C]Alanine (6 GBq/mmol) was purchased from Amersham International plc (Amersham, UK). Preparation qf madijied amino acids and peptides N-Acetylation of amino acids and peptides was performed in 1 :1 mixtures of acetic anhydride and acetic acid at 22 "C for 1 h, followed by evaporation in a centrifugal evaporator or by lyophilisation [lo, 111. N-Formylation of amino acids was similarly performed in 3 : 1 mixtures of formic acid/acetic anhydride at 22 C for 1 h [12, 131. Methyl esters of amino acids were prepared in 0.1 M methanol/HCl, at 22°C for 16 h and the products dried in vucuo [14]. Preparation q f A P H and A P H linked to Sepharose and treated with ~liw?ethylsubrrimidate( D A P H - S ) APH was purified from ovine liver as previously described [6, 151. The enzyme was then coupled to CNBr-activated Sepharose and cross-linked with dimethylsuberimidate as described elsewhere [15]. This preparation is known as DAPH-S. Reverse-phase HPLC All HPLC separations were performed on the System Gold (Beckman, Palo Alto. CA, USA) with a CI8 reverse-phase column (220 mm = 2.1 mm, 5 pm, supplied by Applied Biosystems, Warrington. UK). The flow rate was 0.3 ml/min and the mobile phase was 0.1% aqueous trifluoroacetic acid. Gradients of CHJCN were used as indicated.
For amino acid analysis samples were hydrolysed with 6 M HC1, in the vapour phase, at 110"C for 18 - 20 h, before analysis by an Applied Biosystems 420A amino acid analyzer with on-line model 130A PTC analyser. Protein sequences were determined by automated Edman degradation in an Applied Biosystems 477A protein sequencer with an on-line phenylthiohydantoin amino acid analyser. Fast-atom-bombardment mass spectroscopy ( F A B - M S ) FAB-MS was performed in the positive-ion mode in a VG 250-70 SE instrument, using a 30-keV caesium ion gun. Thioglycerol plus 1% trifluoroacetic acid was used as matrix material. Stimulation of superoxide generation by granulocytes Human granulocytes were purified from 30 ml fresh peripheral blood by overlaying the diluted blood on to layered cushions of Histopaque 1077 and Histopaque 1119 followed by centrifugation at 700 x g for 30 min (according to the manufacturer's instructions). The cells were taken from the interface, washed and used at lo6 viable cells/l ml sample (by trypan blue exclusion) in Krebs-Ringer phosphate solution. Stimulation of superoxide release was measured from the superoxide-dismutase-inhibitable reduction of cytochrome c, as described by Greenberger et al. [16]. Each reaction was performed for 20 min at 37°C in the presence of 0.05 mM cytochrome c and the presence or absence of 0.01 mg/ml superoxide dismutase. The absorbances of the supernatants were meaured at 550 nm. The background absorbances from (a) corresponding superoxide dismutase containing samples, and (b) the negative control with no peptide were subtracted. Hence superoxide production was calculated by dividing by the molar extinction coefficient for the reduction of cytochrome c (21 000 M-'). RESULTS Optimization of conditions f o r synthesis of Ac-Ala-Ala-Ala Ac-Ala-Ala-Ala is known to be a very good substrate for acylpeptide hydrolase [7 - 91. The conditions required for peptide synthesis were therefore initially characterized for the following reaction : Ac-[14C]Ala-OMe MeOH.
+
Ala-Ala
Ac-[14C]Ala-Ala-Ala
+
Use of a large excess of nonradioactive Ala-Ala allowed the radioactive peptide product to be detected with high sensitivity. A time-course of the reaction at 22 "C in aqueous solution is shown in Fig. 2. Hydrolysis to A ~ - [ ' ~ c ] A lwas a the predominant reaction at all time points. Synthesis of Ac['4C]Ala-Ala-Ala reached a plateau at 30 - 60 min, after which synthesis is presumably exceeded by secondary enzymic hydrolysis to Ac-[14C]Ala + Ala-Ala. At no point did the yield exceed 5% (of radioactive component), and at 30 min the amount of Ac-[14C]Ala-Ala-Ala generated was only 16% of the yield of the unwanted product, Ac-[14C]Ala. In order to improve the yield, the effects of organic cosolvents were studied. We have already shown that APH retains some activity in low concentrations (5%, by vol.) of N, N-dimethylformamide, and that N,N-dimethylformamide
689
f 11 .-
m
c c
I
0
I
5
I
15
I
30
Time (min)
I 60
are increased accordingly. Further studies were therefore performed in the presence of unlabelled Ac-Ala-OMe (1 mM). These demonstrated that enzyme action could be accelerated at 45 "C with a corresponding improvement in yield, and that the ratio of A~-['~C]Ala-Ala-Ala/Ac-[~~C]Ala generated was further increased by emulsifying the solution (in the presence of a nonionic detergent) with an organic phase of benzene (not shown). In the latter case, it was subsequently found that Ac-Ala-Ala-Ala did not significantly partition into the benzene phase, so the effect was probably due to the partial solubility of benzene in the aqueous phase, acting like N,Ndimethylformamide to decrease the polarity and H 2 0 activity of the reaction medium. Production of A~-[~~C]Ala-Ala-Ala could thereby be increased to 32% of Ac-[14C]Ala. In this respect, octane (less polar than benzene) had less effect whereas butanol (more soluble in water) inhibited the enzyme.
Fig. 2. Time course of the production of Ac-['4C]Ala-Ala-Ala and Ac[14C]Ala. The reaction mixture contained 1 nmol A c - [ ' ~ C ] A I ~ - O M ~ (3 x lo5 cpm), 50 pmol Ala-Ala and 2 p1 DAPH-S (60 ng APH) in 0.3 ml50 mM NaH2P04/Na2HP04,pH 8, buffer and was incubated Enzymatic synthesis of Ac-Ala-Alu-Ala at 22°C. At the times indicated, 0.05 ml samples were taken, mixed with 0.4 ml ice-cold 0.1% trifluoroacetic acid and briefly centrifuged A larger scale synthesis of Ac-Ala-Ala-Ala was attempted (10 s, 16000 xg). The supernatants were concentrated by freeze-dry- using conditions derived from the preceding results. The reacing then analysed by reverse-phase HPLC in aqueous 0.1% tion mixture (RI) contained the following components : Actrifluoroacetic acid (isocratic conditions) as described in Materials Ala-OMe, 0.5 pmol; Ala-Ala, 50 pmol ; N,N-dimethyland Methods. Fractions corresponding to A ~ - [ ' ~ c ] A l(5 a -6 min, formamide, 5% (by vol.); Nonidet P40, 1% (by vol.); W ) and A~-['~C]Ala-Ala-Ala(11 -13 min, 0 ) were collected and DAPH-S (enzyme), 10 p1 (0.3 pg APH). All these were disradioactivity measured
solved in 0.25 ml 50 mM NaH2P04/Na2HP04,pH 8, mixed with 0.5 ml benzene. A negative control (RO) was included 20 with identical conditions except for the absence of enzyme (DAPH-S). After shaking for 2 h at 45"C, DAPH-S was removed by centrifugation and the supernatant added to 0.5 ml 0.1% (by vol.) trifluoroacetic acid. This inactivated any residual enzyme. Both samples were dried in vucuo and the 15products separated by reverse-phase HPLC (as described in Materials and Methods). The fractions corresponding to the a 10elution position (11- 13 min) of Ac-Ala-Ala-Ala were col8 lected, freeze-dried in four equal aliquots and analysed as a a follows. Q, 5HPLC. One aliquot of each sample was dissolved in 0.1 % a trifluoroacetic acid and refractionated by HPLC as above. R1 contained a peak that co-eluted with an Ac-Ala-Ala-Ala standard and was absent in RO (not shown). This peak had I I I I I been partially obscured by overlap with the much larger peak 0 5 10 20 30 Dimethyl Formamide (%) of Ala-Ala in the original HPLC fractionation of the products, Fig. 3. Effect of",N-dimethylformamide on synthesis of A c - [ ' ~ C ] A I ~ - but was totally resolved in the second chromatography. Alu-Alu. The reaction mixture contained 0.1 nmol A c - [ ' ~ C ] A I ~ - O M ~ FAB-MS. A portion of each sample was analysed by FABMS, as shown in Fig. 4.The predicted ion mass of 274, [Ac(60000 cpm), 10 pmol Ala-Ala, 2 pl DAPH-S (60 ng APH) and the Ala-Ala-Ala HIf in the positive-ion mode, was evident in indicated concentrations of N,N-dimethylformamide in 0.1 ml50 mM NaH2P0,/Na2HP04, pH 8, buffer. After 60 min at 2 2 T , 0.5 ml the sample incubated with enzyme but negligible in the nega0.1% trifluoroacetic acid was added to each. The samples were then tive control. This confirmed that the correct product had been centrifuged, freeze-dried and analysed by HPLC as described in Fig. 2. produced. The generation of A~-['~C]Ala-Ala-Ala is shown as a product ratio Amino acid analysis. One of the aliquots of each sample (a percentage (0)and as total percentage yield ( 0 ) quarter of the total) was subjected to amino acid analysis as described in Materials and Methods. The results indicated 10.33 nmol Ala in R1 compared to 2.98 nmol in RO (alanine is present in RO after HPLC due to overlap with the peak resistance is enhanced in the modified enzyme, DAPH-S [15]. of Ala-Ala). No other residues were present in significant The effect of N,N-dimethylformamide on Ac-[14C]Ala-Alaamounts. The total yield of Ac-Ala-Ala-Ala was therefore Ala synthesis by DAPH-S is shown in Fig. 3. The yield was calculated as 4 x (10.33 - 2.98)/3 = 9.8 nmol. This represents improved in 5% N,N-dimethylformamide, but reduced in a yield of 2% of the Ac-Ala-OMe substrate. higher concentrations, presumably due to inactivation of the enzyme. More importantly for the objective of scaling up the process, 5% N,N-dimethylformamide improved the product Enzymatic synthesis o jfMet-Leu-Phe ratio of Ac-['4C]Ala-Ala-Ala/Ac-[14C]Ala by almost fourfold. This peptide was selected for enzymatic synthesis for the The results predict that significant amounts of product should be obtainable if the amounts of substrate and enzyme following reasons : the formyl group can be removed by mild
d
5-
v
,J
+
acid hydrolysis allowing sequence analysis or addition of further residues; M e t is a favourable terminal group for APH action [6, 91; Met-Leu-Phe has a biological activity, being a potent example of the bacterial formylpeptides that stimulate chemotaxis and trigger a respiratory burst in mammalian granulocytes [17]. The reaction mixture (MLF1) contained the following components: met-OMe, 12.5 pmol; Leu-Phe, 17 pmol; N , N dimethylformamide, 5% (by vol.); Nonidet P40,8% (by vol.); DAPH-S (enzyme), 10 p1 (0.3 pg APH). All were dissolved in 0.25 ml 50 mM NaH2P04/Na2HP04, pH 8, mixed with 0.5 ml benzene. The reaction conditions were as described above for Ac-Ala-Ala-Ala, including the incorporation of a negative control without DAPH-S (MLFO). The products were fractionated by reverse-phase HPLC, as before, but with a 40-min gradient of 0 - 50% CH,CN (started after 10 min in aqueous 0.1 YOtrifluoroacetic acid). The resulting chromatograms are shown in Fig. 5. Only one enzyme-dependent peak is evident (labelled Product) and coincides with the elution volume of standard met-Leu-Phe. This fraction was collected from both samples, freeze-dried in four equal aliquots, and analysed as follows. FAB-MS. FAB-MS, shown in Fig. 6, identified an ion mass of 438 as predicted for [Wet-Leu-Phe HI+ in the positive-ion mode. This species was absent in the negative control. Amino acid analysis. One aliquot of each was subjected to amino acid analysis, yielding the results shown in Table 1. These clearly show that the product contains approximately 2 nmol each of Met, Leu and Phe, indicating a total yield of 8 nmol. This is a 0.064% yield from the substrate Ac-MetOMe. Conversion to Met-Leu-Phe. The formyl group was hydrolysed from one aliquot of each sample in 0.5 M methanol/HCl (24 h at 22°C). A portion of commercial met-Leu-Phe was identically treated in parallel. After drying in vacuo these samples were de-esterified with 0.25 M NaOH (22"C, 1 h) followed by neutralisation with 0.25 M HCl. The products were analysed by reverse-phase HPLC as above. The chromatograms, shown in Fig. 7, indicate that the product now eluted at 44 min (compared to 48 min before hydrolysis, see Fig. 5 ) , with no comparable product in the negative control. This peak was collected and used in sequence analysis as
+APH
100-
274 = IAcAAA+Hl+
1
50-
I
-APH
+
0-I 100
L
200
250
300
400
350
MIZ
Fig. 4. Mass spectrurn of the product of Ac-Ah-Ala-Ah synthesis. FAB-MS in the positive-ion mode was performed on the purified products of the synthesis of Ac-Ala-Ala-Ala under the conditions given in Materials and Methods. The mass ion at 274 corresponds to the m / z (M/Z) predicted for [Ac-Ala-Ala-Ala + HIi, and the ion at 296 corresponds to the sodium adduct [Ac-Ala-Ala-Ala + Na] . The spectrum of the negative control sample without enzyme (-APH) is also shown. The peak at mjz 381 is derived from the thioglycerol matrix and is present in both samples +
I
I 20
I
I
I
I
I
I
I
I
I 60
Time (rnin)
Fig. 5. Sepurution of the products of fMet-Leu-Phe synthesis by reverse-phase HPLC. The reaction conditions for the synthesis of met-LeuPhe are reported in the text. The HPLC conditions are as described in Materials and Methods using a 40-min gradient to 50% (by vol.) CH,CN/O.l% trifluoroacetic acid started after running for 10 min in 0.1 % trifluoroacetic acid. The chromatogram of the negative control sample (-APH) is also shown. The eluate at the position labelled Product (48 min) was collected from both separations for further analysis
69 1 438
=
[fMLF
+
HI+
"'1
+APH (MLF1)
I
a
I
I
I
I
I
I
I
I
I
I B
-0.022
C
1'291 For-MLF
h i l l I
1 -APH (MLFO)
nn Y."
-
I
0
lo] 50
350
450
400
500
5 50
I
I
I
I
Time (rnin)
I
I
I
I
I
60
Fig. I . HPLC analysis of the product of fMet-Leu-Phe synthesis after mild acid hydrolysis. The samples from fMet-Leu-Phe synthesis (+ APH; A), the negative control (- APH; B) and commercial M e t Leu-Phe (For-MLF; C) were prepared and hydrolysed in methanol/ HCI as described in the text. These were then analysed by reversephase HPLC using the same conditions described in Fig. 5 . Successful conversion of fMet-Leu-Phe to Met-Leu-Phe is indicated by the shift in elution position from 48 rnin (in Fig. 5 ) to 44 rnin
I
300
I
600
MI2
Fig. 6. Mass spectrum of the product offMet-Leu-Phe synthesis. Positive-ion FAB-MS was performed on the purified products of the synthesis of fMet-Leu-Phe (described in the text) under the conditions given in Materials and Methods. The spectrum of the negative control sample (MLFO) is also shown. The ion mass of 438 corresponds to the m / z (M/Z) predicted for [Wet-Len-Phe + HI+. The other prominent signals are present in both spectra
bined with the mass spectral data, this confirms that the correct product has been produced. Activity assay. Biological activity of the products was determined from their ability to stimulate the respiratory burst in human granulocytes (as described in Materials and Methods). The results, shown in Table 3, show that the product has a total biological activity slightly greater than the equivalent amount of commercial fMet-Leu-Phe.
Other synthetic products Table 1. Amino acid analysis of the product of fMet-Leu-Phe synthesis Amino acid analysis was performed on an aliquot of each sample as described in Materials and Methods. All other residues detected were present in less than 0.1 nmol. Cys and Trp were not determined _
_
~ ~~~
Amino acid
Amino acids in fMet-Leu-Phe MLFI
MLFO
Leu Phe Met
2.2 2.1 1.5
0.1 0.1 0.0
Ala Are GIY
0.3 0.2 0.1
0.4 0.0 0.1
described below. The standard fMet-Leu-Phe showed a similar shift in elution position after hydrolysis. Sequence analysis. The HPLC-purified deformylated product was freeze-dried and sequenced by automated Edman degradation. The result, shown in Table 2, gives an unambiguous confirmation that the sequence is Met-Leu-Phe. Com-
The range of potential applications was further illustrated by synthesis of the following two dipeptides. The products were identified by HPLC, mass spectroscopy and amino acid analysis. fMet-Met-OMe. The reaction mixture contained the following components: fMet-OMe, 1 pmol; Met-OMe, 10 pmol; N,N-dimethylformamide, 10% (by vol.); DAPH-S, 50 pl. All were dissolved in 0.25 ml50 mM NaH2P04/Na2HP04,pH 8 (without a benzene phase). After 1 h at 45"C, the supernatant was collected, dried in vucuo and fractionated by reverse-phase HPLC (using the conditions described above for Met-LeuPhe). A peak eluting at 40 rnin that was absent in the negative control (without DAPH-S) was collected, dried and analyzed by mass spectroscopy. An ion mass of 323, corresponding to [fMet-Met-OMe + HI', was observed and was not present in the corresponding fraction from the negative control. fLeu-Phe-0 Me. The reaction mixture contained the following components: fLeu-OMe, 12.5 pmol; Phe-OMe, 12.5 pmol; N,N-dimethylformamide, 5% (by vol.); DAPH-S, 50 pl. All were dissolved in 0.25 ml 50 mM NaH2P04/ Na2HP04, pH 8. After 2 h at 45"C, the supernatant was collected, dried in vucuo and fractionated by reverse-phase HPLC (using the conditions described above for Met-Leu-Phe). A peak eluting at 45 min that was absent in the negative control
Table 2. Amino acid .sequence of the product ofjMet-Leu-Phe synthesis Protein sequence analysis was performed on the products after acid hydrolysis and HPLC purification as described in the text and in Materials and Methods. A single unambiguous sequence was obtained (Met-Leu-Phe) Cycle
Amino acid
Amount pmol
1 2 3
Met Leu Phe
30 35 20
Table 3. Biologicul uctivitj of the product of fMet-Leu-Phe synthesis The sample were assayed for biological activity, as described in Materials and Methods by measuring their ability to stimulate the superoxide generation by human granulocytes. The results shown are the means from triplicates (fSEM), compared with the activity of a standard dilution of commercial Met-Leu-Phe Sample
Amount added
Superoxide generation pmo1/lO6 granulocytes
MLFI (enzymatic synthesis) MLFO (control) fMLF (commercial)
1.25% total 1.25% total 0.1I nmol
4.06 f 0.09 0.83 0.23 3.07
(without DAPH-S) was collected and dried. One half was subjected to amino acid analysis, yielding 75 pmol leucine and 117 pmol phenylalanine that were not present in the equivalent fraction of the negative control. The total yield of Ceu-Phe was therefore approximately 0.1 nmol.
DISCUSSION The application of acylpeptide hydrolase in peptide synthesis has been demonstrated by the detailed characterization of the production of two tripeptides (Ac-Ala-Ala-Ala and Met-Leu-Phe). The synthesis of two dipeptides (met-MetOMe and fLeu-Phe-OMe) is also reported. APH can thus be used to elongate peptide chains by sequential addition to the N-terminus. The enzymatically synthesised products are expected to contain only L-amino acids, although this has not been verified experimentally. Importantly, formylated products can be deprotected by mild acid hydrolysis to allow further elongation of the chain. As yet the yields have been too low to perform more than one cycle of amino acid coupling. There are more potential applications of the method than total peptide synthesis. APH could be used to modify preexisting peptides by, for example, introducing a reactive sidechain (e.g. cysteine) or a radiolabel. APH can also be used to remove acetylated or formylated N-terminal residues [5- 9, 151. These amino acids could then be mutated by using the same enzyme to remove residues and, under synthetic conditions, to add back others. It is also possible that APH could couple certain artificial amino acids ( e g homoserine and norvaline) which are difficult to introduce into proteins by genetic means.
The method does have a number of limitations. Firstly, the yields obtained have been poor. This could be improved simply by using much more enzyme but further development is also necessary. Secondly, although the enzyme has quite a broad substrate specificity, it is known not to recognise charged residues at the PI position [8, 9, 181. In total peptide synthesis APH may therefore have to be complemented by other enzymic or chemical methods. Finally the activity of APH declines sharply with increasing size of the substrate [9]. Perhaps this could be overcome by disrupting the secondary structure of the substrate, for example with detergents. The stability of the enzyme would have to be increased which would also allow use of higher concentrations of organic cosolvents. This would improve the yield by reducing the formation of the unwanted hydrolysis product. If the water content could be reduced to a minute fraction it would reduce the ionization of normally charged residues which might then become compatible with enzyme activity. A more stable enzyme might be obtained by chemically modification, genetic engineering or isolation from a thermophilic microorganism [19]. Acylpeptide hydrolase could then become a very useful tool in peptide chemistry. We are grateful to Alan Harris (NIMR, London) for the protein sequence analysis and to Fatima Beg (NIMR) for performing the FAB-MS and amino acid analysis. The financial support of Pharmacia LKB is also gratefully acknowledged.
REFERENCES 1. Merrifield, R. B. (1963) 1. Am. Chem. Soc. 85,2149-2154. 2. Atherton, E., Fox, H., Harkiss, D., Logan, C. J.. Sheppard, R. C. & Williams, B. J. (1978) J . Chem. Soc. Chem. Commun. 537 - 540. 3. Kullmann, W. (1985) J . Protein Chem. 4, 1-22. 4. Jakubke, H.-D. (1987) in Thepeptides. Analysis, synthesis, biology (Udenfriend, S . & Meienhofer, J., eds), vol. 9, pp. 103-165, Academic Press, San Diego. 5. Tsunasawa, S . , Narita, K. & Ogata, K. (1975) J . Biochem. (Tokyo) 77,89-102. 6. Tsunasawa, S . & Narita, K. (1976) Methods Enzymol. 45, 552561. 7. Gade, W. &Brown, J. L. (1978) J . Bid. Chem. 253, 5012-5018. 8. Jones, W. M. & Manning, J. M. (1985) Biochem. Biophys. Res. Cornmun. 126,933-940. 9. Kobayashi, K. & Smith, J . A. (1987) J . Riol. Chem. 262,1143511445. 10. Green, R. W., Ang, K. P. & Lam, L. C. (1953) Biochem. J . 54, 181 - 187. 11. Riordan, J. F. & Vallee, B. L. (1972) Methods Enzymol. 25,494499. 12. Sheehan, J. C. & Yang, D.-D. H. (1957) J . Am. Chem. Soc. 80, 1154- 1158. 13. Geiger, R. & Konig, W. (1981) in Thepeptides. Analysis, sjwthesis, biology (Gross, E. & Meienhofer, J., eds) vol. 3, pp. 1-99, Academic Press, New York. 14. Wilcox, P. E. (1967) Methods Enzymol. 11, 605-617. 15. Farries, T. C., Harris, A,, Auffret, A. D. & Aitken, A . (1991) Eur. J. Biochem. 196, 679-685. 16. Greenberger, J. S . , Newburger, P. E., Karpas, A. & Moloney, W. C. (1978) Cancer Res. 38,3340-3348. 17. Showell, H. J., Freer, R. J., Zigmond, S . H., Schiffmann, E., Aswanikumar, S . , Corcoran, B. & Becker, E. L. (1976) J . Exp. Med. 143,1154-1169. 18. Orfeo, T. & Meyer, W. L. (1986) Fed. Proc. 45, 1712. 19. Mozhaev, V. V., Berezin, I. V. & Martinek, K. (1988) CRC Crit. Rev. Biochem. 23,235-281. 20. Farries, T. C., Auffret, A. D. & Aitken, A. (1990) Swedish patent application No. SE 9000 597-6.
