E L S E V I E R Journal of Biotechnology 32 (1994) 139-148

journal of blotechnology

High-level expression of porcine muscle adenylate kinase in Escherichia coli: Effects of the copy number of the gene

and the translational initiation signals

Takeshi Hibino a,1, Satoru Misawa a,2, Motoaki Wakiyama b, Shu Maeda b, Kazumori Yazaki c, Izumi Kumagai a, Tatsuo Ooi a,3, Kin-ichiro Miura a,,

a Department oflndustrial Chemistry, Faculty of Engineering, University of Tokyo, Hongo, Tokyo 113, Japan b Institute for Biomolecular Science, Faculty of Science, Gakushuin University, Mejiro, Tokyo 171, Japan

c Ultrastructural Research Unit, Tokyo Metropolitan Institute of Medical Science, Honkomagome, Hongo, Tokyo 113, Japan

(Received 11 January 1993; revision accepted 12 March 1993)

Abstract

Porcine muscle adenylate kinase (ADK) was overproduced in Escherichia coli using the expression plasmid with double A-T-G codon at the translational starting site and the Shine-Dalgarno (SD) sequence 10 bp apart from the first A-T-G. We used the expression vectors pKK223-3 and pMK2. pMK2 is about 10 ~ 20-times larger in copy number than pKK223-3. For both vectors, duplication of A-T-G was effective and the quantity of the expressed ADK from the double A-T-G plasmid was 2 ~ 4-fold more than that achieved when only one A-T-G was present. The amount of the produced ADK was maximum in the case of using pMK2 with double A-T-G. The overproduced ADK formed inclusion bodies in E. coli. It was solubilized in 6 M guanidine hydrochloride and refolded. Through two steps of column chromatography, ADK was purified. It has the same amino acid composition and grossly the same activity as that reported by Schirmer et al. (1970). Its amino acid sequence of the NH2-terminal region was identical with that deduced from the cDNA sequence including the NH2-terminal methionine.

Key words: Adenylate kinase; Expression vector; Translation; Initiation codon A-T-G

* Corresponding author (Present address): Institute for Biomolecular Science, Faculty of Science, Gakushuin Univer- sity, 1-5-1 Mejiro, Toshima-ku, Tokyo 171, Japan. Present addresses: 1 Laboratory of Biophysical Chemistry, College of Agriculture, University of Osaka Prefecture, Sakai, Osaka 591, Japan, 2 Pharmaceuticals and Biotechnology Lab- oratory, Central Research Laboratory, Research and Devel- opment Division, Nikko Kyodo Co. Ltd., Toda, Saitama 335, Japan, 3 Department of Food Science, Kyoto Women's Uni- versity, Higashiyama-ku, Kyoto 605, Japan.

I. Introduction

Adenylate kinase is a monomeric enzyme of 20 ~ 30 kDa catalyzing the reaction as described below and regulates the energy supplying cycle in the cell.

M g - A T P + A M P ~ > M g - A D P + A D P

0168-1656/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSDI 0168-1656(93)E0022-P

140 T. Hibino et al. /Journal of Biotechnology 32 (1994) 139-148

Various kinds of ADK have been purified from several species and characterized (Noda and Kuby, 1957; Chiga and Plaut, 1960; Markland and Wadkins, 1966; Sapico et al., 1972; Tsuboi and Chervenka, 1975; Itakura et al., 1978; Tomasselli and Noda, 1980; Ito et al., 1980; Yeh et al., 1983; Batra et al., 1986). Porcine muscle adenylate kin- ase comprises 194 amino acid residues, a single polypeptide chain of 21.4 kDa and has two free cysteines. Its amino acid sequence was reported by Heil et al. (1974) and three-dimensional struc- ture was determined by Schulz et al. (1977), Sach- esenheimer and Schulz (1977) and Pai et al. (1977). ADK seemed to be very suitable for stud- ies on elucidation of nucleotide-recognition mechanism of protein because of the accumula- tion of the information about its function and structure. The cDNA for porcine skeletal muscle ADK have already been cloned (Hibino et al., unpublished data). We report here the markedly high-level expression system of porcine muscle ADK in E. coli and characterization of the pro- duced ADK.

In this study, the cDNA fragment of porcine muscle ADK was introduced into the expression vector pKK223-3 (Brosius and Holy, 1984) locat- ing the initiation codon A-T-G 10 bp downstream of the Shine-Dalgarno (SD) region (Shine and Dalgarno, 1974). The constructed plasmid pK- KAK1-2 expressed detectable amounts of ADK in transformed E. coli. To improve the expression system, modifications were made mainly at two points. Firstly, the replication origin of pKK223-3 was replaced with that of pUC18 (Yanisch-Per- ron et al., 1985) to construct the high copy num- ber vector pMK2. Secondly, the additional A-T-G was inserted just behind the original initiation codon A-T-G. We considered the possibility that the second A-T-G may act as a 'fail-safe' transla- tion initiator. Increasing the copy number of the expression vector gave rise to great enhancement of the ADK expression. Duplication of A-T-G at the translational initiation site was considerably effective for both cases using pKK223-3 or pMK2, and caused the increase of the expression of ADK about 3- to 4-fold and 2-fold more than that achieved when only one A-T-G was present, re- spectively.

2. Materials and methods

2.1. Bacterial strains and plasmids

E. coli strains JM105 [supE endA sbcB15 hsdR4 rpsL thi A(lac-proAB) F'[traD36 proAB + lacI q IacZAM15]] and JM109 [recA1 supE44 endA1 hsdR17 gyrA96 relA1 thi a(lac-proAB) F'[traD36 proAB + lacI q IacZAM15]] were used as the expression hosts. The plasmid pPAK3-1 contains the PstI fragment bearing the coding region of ADK (Hibino et al., unpublished data).

2.2. Construction of the expression plasmid

The PstI fragment of ADK cDNA was di- gested with Bal31 followed by treatment with Klenow enzyme to create blunt end. Each short- ened cDNA was subcloned into SmaI site of pUC18 as shown in Fig. 1. An EcoRI-HindlII fragment, which lost the 5 ' - (dG/dC) tail and had intact 3 ' - (dG/dC) tail, was cloned between EcoRI and HindlII sites of pKK223-3 and pKKAK1 was constructed. Then the nucleotides of the internal region between SD and A-T-G were deleted and the SD-(A-T-G) length was converted to be 10 nucleotides by site-directed mutagenesis (Morinaga et al., 1984) using the mutagenic primer 1 (*) and pKKAK1-2 was con- structed.

(*) Primer 1; 5'CAGGAAACAGAATFCATGG- AAGAGAAGCTG 3'

The multi-copy number expression vector pMK2 was made by replacing the replication origin of pKK223-3 with that of pUC18, pKK223-3 was double digested with NruI /PvuI , and the 1.8 kbp fragment was ligated with the 1.4 kbp PvuI- PvulI fragment of pUC18 containing the region of the replication origin of it. Insertion of the additional A-T-G codon next to the initiation methionine codon was carried out by site-di- rected mutagenesis using the mutagenic primer 2(**)

(* *) Primer 2; 5,CAGGAAACAGAATTCATG_ ATGGAAGAGAAG 3'

T. Hibino et al. /Journal of Biotechnology 32 (1994) 139-148 141

The resulting expression plasmids are called as follows; the pKK223-3 derived one A-T-G and double A-T-G plasmids are called pKKAK1-2 and pKKAK3, the pMK2 derived one A-T-G and double A-T-G plasmids are called pMKAK1-2 and pMKAK3.

2.3. Northern blot hybridization analysis of A D K mRNA

Bulk RNA was prepared from E. coli cells transformed with pKKAK1-2 or pKKAK3 by the hot phenol method (Aiba et al., 1981). Northern RNA-DNA hybridization analysis was performed essentially as described by Sambrook et al. (1989) using the 32p-labeled synthetic oligonucleotide probe

S,GAGCCGCTrGGTCATGGTCTCAGGGCC_ CGCG 3,

corresponding to 357 ~ 387 nucleotides (in the case of pKKAK1-2) of the ADK coding region.

2.4. Purification and refolding of ADK expressed in E. coli

The overnight culture of E. coli JM109 har- boring pMKAK1-2 was harvested by centrifuga- tion, washed with 0.5% NaCI, 0.5% KCI solution and recovered. Then the cells were suspended in 2 vol. of 20 mM Tris-HC1 at pH 7.4, 100 mM NaC1, 1 mM /3-mercaptoethanol, 0.1 mM EDTA and disrupted by sonication. The insoluble mate- rials were recovered from the cell lysate by cen- trifugation at 8000 × g for 20 min. Washed with 5% Triton X-50 solution, the inclusion bodies were recovered by centrifugation at 8000 × g for 10 min. Then the included ADK was solubilized in 6 M guanidine-hydrochloride, 20 mM Tris at

Sma I Pvu I I °aI

N r u i ~ ~ EcoRI~ ~_ HindlII

Ec:RI ECCn~Iii

S/TaI /T5 DNA Ligase Smal ~ " EcORI ~Imb~ HindI I I

EcoRI/~ ~HindIII pt ac x'./~ ~ ~'~k~ I

pKKAK i pKKAK 1 -

A / Site-directed ~ / .ut..enes~s

pKK223-3

I vui NruI

ECoRI HindIII

Ptacj~.~ PvuII NruI~/

PVUI

pUC18

~ PVUI PVUI I

i. 4Kbp

\ . ..J, oI ori

$ T4 DNA Ligase EcoRI HindIII Ptac

~ P v u l

ori

Fig. 1. Construction strategy of the expression plasmid pKKAK1-2 and the high-copy number expression vector pMK2.

142 T. Hibino et al. /Journal of Biotechnology 32 (1994) 139-148

pH 7.4, 0.1 mM EDTA. The solution containing ADK was dialyzed against 4 vol. of Tris buffer overnight. The buffer were changed three times after every 4 ~ 6 h during dialysis. After dialysis, the solution containing ADK was charged on Affigel Blue colummn (15 x 1.2 cm) pre-equi- librated with 20 mM MOPS at pH 7.2, 100 mM NaCI, 1 mM DT-I', 0.1 mM EDTA. Then the column was washed with 5 column volume of MOPS buffer, and ADK was eluted out with MOPS buffer by applying the linear gradient of NaC1 concentration of 0.1 M to 3 M. Every frac- tion of Affigel chromatography was monitored by 280 nm ultraviolet absorption and several of them were assayed with P K / L D H coupled ADK assay system (Agarwal et al., 1979) at the substrate concentration of 100/xM of both ATP and AMP. The fractions containing ADK were combined and concentrated by ultrafiltration on Amicon concentration apparatus. The concentrated ADK solution was charged on Sephacryl S-200 column (1.2 x 120 cm) pre-equilibrated with Tris buffer and eluted out with the same buffer, all the fractions were monitored by its 280 nm ultraviolet absorption and several fractions were also as- sayed with the ADK activity at the same condi- tions as above. The fractions containing ADK were combined and used for further characteriza- tion. The purity of the isolated ADK was checked by SDS-polyacrylamide gel electrophoresis (Laemmli, 1970).

2.5. Amino terminal analysis and amino acid com- position analysis

Some portion of the ADK purified as de- scribed previously was further purified by HPLC with C18 column in 0.1% trifluoroacetic acid from 0 to 80% of acetonitrile gradient. Authentic ADK isolated from porcine muscle was also puri- fied on C18 column under the same conditions. The HPLC-purified ADK was lyophilized. Two /~g of them were sequenced starting from the NH2-terminal using Applied Biosystems gas- phase protein sequencer model 470 A. Two hun- dred /~g of them was hydrolyzed in 6 N HC1 at 120°C for 8 h and the hydrolysates were subjected to the analysis of the amino acid composition

using the amino acid analyzer, Beckman system 7300 high performance analyzer. The enzyme ac- tivity was measured by P K / L D H coupled assay system (Agarwal et al., 1979) using Mg-ATP and AMP concentration of 400/~M each. The enzyme concentration was estimated by the ultraviolet absorption at 280 nm of the ADK solution (E 1% at 280 = 6.0 (Kalbitzer et al., 1983)).

3. Results

3.1. Construction of expression plasmid of porcine ADK

The EcoRI-HindlII fragment of cDNA was introduced into EcoRI-HindlII sites of pKK223-3. This construction gave no expression product of the introduced gene. Since the long distance be- tween A-T-G and SD (90 bp apart) would pre- vent ADK from being translated, the SD-(A-T-G) distance was converted to be 10 bp. The sequence between SD and A-T-G was determined in such a way that d G / d C base pairs were selectively deleted and d A / d T base pairs were selectively introduced to prevent potential secondary struc- ture formation. However, the resulting plasmid pKKAK1-2 expressed ADK with rather poor effi- ciency, less than 1% of the total cellular protein. The immunoblotted ADK was identified as a unique band for the E. coli lysate harboring pKKAK1-2 at the correspondent position to the ADK from porcine muscle (data not shown). Such protein band was not detected with the lysate of E. coli which did not carry the expression plas- mid. The SDS-PAGE gel also showed a minute band at the position corresponding to the immun- odetected ADK. We made other 3 types of ex- pression plasmids which were different in length from 7 to 12 bp of the SD-(A-T-G). Efficiency of the expression of ADK was compared to pKKAK1-2 on SDS-PAGE. Only slight difference was detected by densitometry (data not shown).

3.2. High level expression system of porcine ADK in E. coli

To overproduce porcine ADK in E. coli, the copy number of the expression vector was in-

T. Hibino et al. /Journal of Biotechnology 32 (1994) 139-148 143

creased by replacing the replication origin of pKK223-3 with that of pUC18. The copy number of the resulting plasmid pMK2 was found to increase 10- to 20-times in comparison with that of pKK223-3 on agarose gel electrophoresis scan- ning by densitometry (data not shown). The high-copy number expression plasmid pMKAK1-2 produced ADK with great efficiency. Highly ex- pressed ADK formed inclusion bodies in E. coli. Fig. 2 shows the electron microscopic picture of inclusion body formed in elongated E. coli JM105 harboring pMKAK1-2. The inclusion body showed variety of size from 0.1 to 2.0 ~m in diameter.

Addition of another A-T-G just behind the original initiation codon A-T-G gave considerable

effect on the expression of A D K (Fig. 3). The expression efficiency was raised about 3- to 4-fold for pKKAK3 and 2-fold for pMKAK3 in compar- ison with pKKAK1-2 and pMKAK1-2, respec- tively. The amount of the produced ADK was maximum in the case of using pMKAK3 without IPTG induction (Fig. 4). IPTG-induced E. coli harboring pMKAK3 duplicated slowly and pro- duced small amount of ADK.

3.3. Northern blot hybridization analysis o f A D K m R N A

The steady state level of ADK mRNA was examined by Northern blot hybridization analysis (Fig. 5). The wide and the thin arrow indicates

Fig. 2. Electron microscopic photograph of the inclusion bodies containing ADK formed in E. coli JM105 harboring pMKAK1-2. Overnight cultured E. coli cells were harvested and washed with 1% NaCI and stained with 1% uranic acetate. (A) E. coli expressed ADK as inclusion bodies. The dense material shown in the elongated E. coli seemed to be inclusion bodies. (B) E. coli not expressed ADK.

144 T. Hibino et al. /Journal of Biotechnology 32 (1994) 139-148

the intact and the degraded ADK mRNA, re- spectively. Quantitation of the RNA products by densitometric scannning of the autoradiogram showed that the level of the intact ADK mRNA transcribed from pKKAK3 was 2.4-times larger than that transcribed from pKKAK1-2 in the case that the E. coli cell was induced with IPTG. In addition, it is noticed that the fraction of the intact mRNA was larger than that of the de- graded mRNA for pKKAK3 but smaller for pKKAK1-2.

3.4. Preparation and characterization of A D K ex- pressed in E. coli

E. coli cells were disrupted by sonication and the inclusion bodies were selectively precipitated by centrifugation. Refolding of ADK was carried

[ a ] 10bp A ~

pKIe~.K1-2 ~Pt-a-~--~l . . . . . . . . . ~ATGGAA

pKKAK3 ~]--[-S-~ ............ ATGATGGAA

ADK

Fig. 4. SDS-PAGE analysis of the proteins in the extract of E. coli harboring pMKAK3. The arrow indicates the position of ADK.

[bl 1 2 3 4

AD K

Fig. 3. (a) Construction of pKKAK1-2 and pKKAK3. (b) SDS-PAGE analysis of the extract of E. coli harboring pKKAK1-2 or pKKAK3. The gel was stained with Coomassie brilliant blue. Lanes 1 and 2: total protein extract from E. coli harboring pKKAK1-2 and p KAK3, respectively. Lanes 3 and 4: total protein extract from E. coli harboring pKKAK1-2 and pKKAK3 induced with IPTG, respectively. The arrow indi- cates the position of ADK.

out by dialyzing the solubilized ADK solution against Tris-HC1 buffer at 4°C for overnight. Fig. 6 shows the chromatogram of the refolded ADK on Affigel Blue column. The fractions indicated by the horizontal bar above the figure were com- bined and used for further purification. Fig. 7 shows the chromatogram of Affigel Blue-puri- fied-ADK on Sephacryl S-200 gel filtration col- umn. The fractions indicated by the horizontal bar were combined and subjected to characteriza- tion.

The amino acid compositions of the purified ADK are listed in Table 1 in comparison with that of the authentic porcine ADK. The ADK expressed in E. coli had an identical amino acid composition to that reported one (Hell et al., 1974). The sequence of the 20 amino acid residues of the ADK expressed in E. coli was determined from the NH 2-terminus. The NH z-terminal amino

T. Hibino et aL /Journal of Biotechnology 32 (1994) 139-148 145

acid sequence of AK1-2 was de t e rmined to be M e t - G l u - G l u - L y s - L e u - L y s - L y s - S e r - L y s - I l e - I l e - Phe-Val-Val-Gly-Gly-Pro-Gly-Ser-Gly, which was the same as that deduced from the c D N A se- quence including the NH2- te rmina l Met, corre- sponding to the ini t ia t ion codon. In addit ion, AK-3 was conf i rmed to have t a n d e m repea ted me th ion ine at its NH2- terminus .

A

0.6

ol0.4

"~o.2

0.0

t i

.. 3.0-

~ .I- :,, 1.0 ~ 2.0-

° l J J " a [YO.5 . " ' I 0 .~ / , \, . -

-}' ~0.I , , - W..

L - ' " " " • • 0.0

A 2b 4~ 6b Bb ,6o ,~o Fruction Number

o ~ u

1 2 3 4 A 10 20 30 40 50 60 70 80 90 100 110 120 130

-'- 23S

--- 16S

i n t a c t

degraded

B

3 0 0

~I p g g A g l - 2 [] pXXXX3

,-4 2 0 0

,,,4

~ 1 0 0

0 i n t a c t d e g r a d e d

Fig. 5. Northern blot hybridization analysis of ADK mRNAs transcribed from pKKAK1-2 and pKKAK3. (A) Lanes 1, 2: Bulk RNA from E. coli harboring pKKAK1-2 (1) and pKKAK3 (2). Lanes 3 and 4: Bulk RNA from E. coli harbor- ing pKKAK1-2 (3) and pKKAK3 (4) induced with IPTG. The wide and the thin arrow indicates the intact and the degraded ADK mRNA, respectively. The positions of the 16S and the 23S rRNA are marked. (B) Densitometric scanning data of the autoradiogram of lanes 3 and 4.

5

Fig. 6. Affigel Blue chromatogram of the produced ADK from pMKAK1-2. (A) Refolded ADK was charged on Affigel Blue column (15 × 1.2 cm) and eluted in MOPS buffer (MOPS 50 mM Tris-HCl pH 7.2, 1 mM fl-mercaptoethanol, 0.1 mM EDTA) by linear concentration gradient of 0.1 M to 2 M NaC1. A28 o was plotted ( ) and the enzyme activity of several fractions were also shown on the plot ( . . . . . . ). The ADK activity was measured by the pyruvate kinase-lactate dehydrogenase-coupled assay system. The fractions indicated with a horizontal bar were combined as an active ADK. (B) SDS-PAGE analysis of the several fractions. Lane A: ADK purified from porcine muscle. Fraction numbers were marked on the picture.

The enzyme activity de te rmined at ATP, A M P concen t ra t ion of 400 ~ M each with enzyme cou- pled assay was almost the same as repor ted value (Schirmer et al., 1970). Above that concent ra t ion , the enzyme showed the substrate A T P inhibi t ion and the velocity slowed down.

4. Discussion

In this paper we showed the markedly high- level expression system of the porcine muscle

146 T. Hibino et al. /Journal o f Biotechnology 32 (1994) 139-148

ADK in E. coli. The amount of the produced ADK was maximum in the case of using pMKAK3 without IPTG induction. IPTG-induced E. coli harboring pMKAK3 duplicated slowly and pro- duced small amount of ADK. The grown E. coli cells still had the plasmid. Too much expression of ADK may be harmful for the E. coli cells, and the expression system may be damaged.

Raising the efficiency of the expression with increasing of the plasmid copy number is quite reasonable. But the level of the expressed ADK

1.5

o 1.0

.~ o.s

0.0

uri~slpl

-12.o ~

Ioo A 2b ~b ~b 86 ,0'0

Frtaclion Number

Table 1 Amino acid composition of porcine muscle ADK expressed in E. coil and from porcine skeletal muscle

No. ADK from Recombinant porcine muscle ADK

res % res %

Asx 13 13.5 7.7 13.1 7.3 Thr 14 12.2 7.0 12.6 7.0 Ser 11 9.3 5.3 9.5 5.3 Glx 25 25.0 14.3 25.0 14.0 Gly 19 17.4 9.9 18.7 10.4 Ala 8 9.1 5.2 8.4 4.7 Val 17 15.1 8.6 15.6 8.6 Met 6 5.3 3.0 5.5 3.1 Ile 9 7.6 4.3 7.6 4.2 Leu 18 18.0 10.3 18.1 10.1 Tyr 7 5.3 3.0 5.1 2.8 Pbe 5 5.3 3.0 5.0 2.8 His 2 2.3 1.3 3.0 1.7 Lys 21 20.2 11.6 22.2 12.3 Arg 11 9.8 5.6 10.7 5.9 Pro 6 6.5 3.6 6.2 3.3

HPLC purified ADK was hydrolyzed for 8 h at 120°C in 6 N HCI under vacuum and then subjected to amino acid analysis. The number of cysteine residues could not be determined by improper experimental conditions. Reported residue numbers of muscle ADK are shown in the leftmost column (No.). The deduced residue numbers (res) and the percent composition (%) are shown for each sample.

A o50 *60 o70 o80 o90

8

Fig. 7. Sephacryl S-200 chromatogram of ADK. (A) ADK purified by Affigel Blue column chromatography was concen- trated and further purified by Sephacryl S-200 column chro- matography (1.1× 140 cm). Absorbance A280 and the ADK activity of the fractions were plotted. The fractions indicated with a horizontal bar were combined as an active ADK. (B) SDS-PAGE analysis of several fractions. Lane A: ADK puri- fied from porcine muscle. Fraction numbers were marked on the picture.

protein increased 2- to 3-times more than the increase of the copy number of the expression plasmid. The increase in the expression would be caused not only by the increase in the mRNA quantity in consequence of the increase in the copy number of the expression plasmid, but also by some other factors; for example, the increase in the stability of the mRNA. This was suggested from the analysis of the expression system using pKK223-3 vector.

The amount of the expressed ADK from pKKAK3 was about 3- to 4-fold that from pKKAK1-2, and the steady state level of mRNA transcribed from pKKAK3 was 2.4-fold larger than that transcribed from pKKAK1-2. It is un- likely that the alterations around the initiation codon A-T-G affect the frequency of the tran- scriptional initiation, because the transcriptional starting site is more than 30 bases upstream of the first A-T-G. One explanation for the ob-

T. Hibino et aL /Journal of Biotechnology 32 (1994) 139-148 147

served differences could be a different rate of degradation. Insertion of another A-T-G just be- hind the original initiaton codon A-T-G may in- crease the efficiency of binding of the ADK mRNA to ribosome, resulting in the increase of the stability of mRNAs. This is in good agree- ment with the fact that the fraction of the intact mRNA was larger for pKKAK3 than pKKAK1-2. It was suggested that the base sequences near the initiation codon A-U-G affected the interaction of mRNA with ribosome (Tsuda et al., 1987). Looman et al. (1987) reported about the influ- ence of the codon following the initiation codon A-T-G on the expression of a modified lacZ gene in E. coli. In their case, G-A-A(Glu) was better than A-T-G(Met) for the second codon. This result does not hold for our data. Although the mechanism of the enhancement of ADK expres- sion by the duplication of the initiation codon A-T-G has not been clear, the enhancement would occur at the translational level. In the case of the expression plasmid carrying duplicated A- T-G, whether the first A-T-G is the unique initia- tor or not is an important question. The NH 2- terminal amino acid sequence of porcine ADK expressed in E. coli was identified to be Met Glu--- for AK1-2 and Met Met Glu--- for AK3. Consequently, in the case of the double A-T-G plasmid, the first A-T-G seemed to be recognized as the initiaton codon. The second A-T-G may have a role in reinforcing the interaction between the ribosome and the mRNA at the initial step.

5. References

Agarwal, K.C., Miech, R.P. and Parks, R.E. (1979) Guanylate kinase from human erythrocytes, hog brain and rat liver. Methods Enzymol. 5l, 483-490.

Aiba, H., Adhya, S. and De Chrombrugghe, B. (1981) Evi- dence for two functional gal promoters in intact Es- cherichia coli cells. J. Biol. Chem. 256, 11905-11910.

Batra, P.P., Burnette, B. and Takeda, K. (1986) Purification and characterization of ATP:AMP phosphotransferase from Mycobacterium marinum. Biochim. Biophys. Acta 869, 350-357.

Brosius, J. and Holy, A. (1984) Regulation of ribosomal RNA promoters with a synthetic lac operator. Proc. Natl. Acad. Sci. USA 81, 6929-6933.

Chiga, M. and Plaut, G.W. (1960) Nucleotide transphosphory- lases from liver. J. Biol. Chem. 235, 3260-3265.

Heil, A., Muller, G., Noda, L., Pinder, T., Schirmer, H., Schirmer, I. and Von Zabern, I. (1974) The amino-acid sequence of porcine adenylate kinase from skeletal mus- cle. Eur. J. Biochem. 43, 131-144.

Itakura, T., Watanabe, K., Shiokawa, H. and Kubo, S. (1978) Purification and characterization of acidic adenylate ki- nase in porcine heart. Eur. J. Biochem. 82, 431-437.

Ito, Y., Tomasselli, A.G. and Noda, L.H. (1980) ATP: AMP phosphotransferase from Baker's Yeast. Eur. J. Biochem. 105, 85-92.

Kalbitzer, H.R., Marquetant, R., Connolly, B.A. and Goody, R.S. (1983) Structural investigations of the Mg-ATP com- plex at the active site of porcine adenylate kinase using phosphorothioate analogs and electron paramagnetic reso- nance of Mn(I1) with chiral 1TO-labelled ATP analogs. Eur. J. Biochem. 133, 221-227.

Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685.

Looman, A.C., Bodlaender, J., Comstock, L.J., Eaton, D., Jhurani, P., De Boer, H.A. and Van Knippenberg, P.H. (1987) Influence of the codon following the AUG initia- tion codon on the expression of a modified lacZ gene in Escherichia coli. EMBO J. 6, 2489-2492.

Markland, F.S. and Wadkins, C.L. (1966) Adenosine triphos- phate-adenosine 5'-monophosphate phosphotranspherase of bovine liver mitochondria. J. Biol. Chem. 241, 4124- 4135.

Morinaga, Y., Franceschini, T., Inoue, S. and Inoue, M. (1984) Improvement of oligonucleotide-directed site-spe- cific mutagenesis using double-stranded plasmid DNA. Bio/Technology 2, 636-639.

Noda, L. and Kuby, S.A. (1957) Adenosine triphosphate- adenosine monophosphate transphosphorylase (myoki- nase): I. Isolation of the crystalline enzyme from rabbit skeletal muscle. J. Biol. Chem. 226, 541-549.

Pal, E.F., Sachesenheimer, W., Schirmer, R.H. and Schulz, G.E. (1977) Substrate positions and induced-fit in crys- talline adenylate kinase. J. Mol. Biol. 114, 37-45.

Sachesenheimer, W. and Schulz, G.E. (1977) Two conforma- tions of crystalline adenylate kinase. J. Mol. Biol. 114, 23-36.

Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular Cloning, Cold Spring Harbor Laboratory Press, New York, USA.

Sapico, V., Litwack, G. and Criss, W.E. (1972) Purification of rat liver adenylate kinase isozyme II and comparison with isozyme III. Biochim. Biophys. Acta 258, 436-445.

Schirmer, I., Schirmer, R.H., Schulz, G.E. and Thuma, E. (1970) Purification, characterization and crystallization of pork myokinase. FEBS Lett. 10, 333-338.

Schulz, G.E., Elzinga, M., Marx, F. and Schirmer, R.H. (1977) Three-dimensional structure of adenyl kinase. Nature 250, 120-123.

148 T. Hibino et al. /Journal of Biotechnology 32 (1994) 139-148

Shine, J. and Dalgarno, L. (1974) The 3'-terminal sequence of Escherichia coli 16S ribosomal RNA: Complementarity to nonsense triplets and ribosome binding site. Proc. Natl. Acad. Sci. USA 71, 1342-1346

Tomasselli, A.G. and Noda, L.H. (1980) Miotochondrial ATP: AMP phosphotransferase from beef heart: purification and properties. Eur. J. Biochem. 103, 481-491.

Tsuboi, K.K. and Chervenka, C.H. (1975) Adenylate kinase of human erythrocyte. J. Biol. Chem. 250, 132-140.

Tsuda, S., Ooi, T. and Miura, K. (1987) Possible interaction

sites of messenger RNA with ribosome deduced from statistical characterization of base sequences. Bull. Inst. Chem. Res. 65, Kyoto Univ., 231-241.

Yanisch-Perron, C., Vieira, J. and Messing, J. (1985) Im- proved M13 phage cloning vectors and host strains: nu- cleotide sequence of the M13mpl8 and pUC19 vectors. Gene 33, 103-119.

Yeh, S.S., Tomasselli, A.G, and Noda, L.H. (1983) ATP-AMP phosphotransferase from Paracoccus denitrificans. Eur. J. Biochem. 136, 523-529.