J. Biochem. 110, 111-119 (1991)

The Nucleotide Sequence for a Proline-Activating Domain of S Synthetase 2 Gene from Bacillus brevis1

Kazuko Hori,* Yoshihiro Yamamoto,* * Kazuhiko Tokita,* Fumihiko Saito,* Toshitsugu Kurotsu, * Masayuki Kanda, * Kaori Okamura, * Junichi Furuyama, and Yoshitaka Saito* *Department of Biochemistry and *'Department of Genetics , Hyogo College of Medicine, Mukogawa-cho, Nishinomiya, Hyogo 663

Received for publication, January 24, 1991

A fragment encoding proline-activating domain (grs 2-pro) of gramicidin S synthetase 2 (GS 2) was found in an 8.1-kilobase pairs (kb) DNA fragment of Bacillus brevis Nagano, which contained the full length of GS 1 gene (grs 1). The clones designated GS719 and GS708, which expressed gramicidin S synthetase 1, were elucidated to express immunoreactive

proteins to GS 2 antibodies with approximate molecular weights of 115,000, 105,000 (GS719), and 110,000 (GS708). The partial purification of the gene products of these clones was carried out using DEAE-Sepharose CL-6B column chromatography. The immunore active proteins to GS 2 antibodies were separated from gramicidin S synthetase 1 protein and had specific proline-dependent ATP-32PP, exchange activity. The nucleotide sequence for the proline-activating domain in the 8.1-kb insert was determined. This fragment was 2,879 base pairs long, and encoded 959 amino acids. The calculated molecular weight of 111,671 was consistent with the apparent molecular weight of 115,000 found in SDS-PAGE of the immunoreactive products to GS 2 antibodies. The open reading frame for this protein followed grs 1 gene, though two were separated by a 73-base pair noncoding sequence, and remained open to the end. A comparison of the putative amino acid sequence of grs 2-pro gene showed that the amino-terminal region was highly homologous to the available sequence of tyrocidine synthetase 2 (71% similarity) and the carboxy-terminal region of the gene product had significant homology to the amino-terminal half of GS 1(66% similarity), as well as to other peptide synthetases, tyrocidine synthetase 1 and ƒÂ-(L-ƒ¿-aminoadipyl)-L-cysteinyl-D-valine synthetase (ACVS), an involved in

biosynthesis.

A cyclic decapeptide antibiotic, gramicidin S, (D-Phe- and included the full length of GS 1 gene. The sequence L-Pro-L-Val-L-Orn-L-Leu)2, is produced by various strains analysis of GS 1 gene was carried out using pGS309 (7). of Bacillus brevis at the beginning of the stationary phase of Three open reading frames of sufficient length to encode the growth. The synthesis of gramicidin S is catalyzed by protein were present in the 4.6-kb insert. GS 1 gene was the gramicidin S synthetase 1 (GS 1) [EC 5.1.1.11 ] and 2 (GS second open reading frame and the others were incomplete 2), which have molecular weights of 120,000 and 280,000, frames. The other clones, pGS708 and pGS719, contained respectively (1). GS 1 activates, thioesterifies, and race longer inserts (8.1kb) which were the same but inserted in mizes phenylalanine (2) and GS 2, a multifunctional opposite orientations (7). These inserts involved two other polypeptide, activates and thioesterifies other constituent open reading frames of about 1.5 and 3kb long, upstream amino acids (3, 4). Peptide elongation is initiated by the and downstream from the grs 1 gene, respectively. transfer of D-phenylalanine from GS 1 to GS 2 (5), subse Recently Kratzschmar et al. reported the DNA sequence quent peptidyltransfers to 4•L- and from of about 5.9 kb of the gramicidin S synthetase gene, the to the following specific amino acid give rise to including the full length of the GS 1 gene and a new gene a pentapeptide (6), and then two identical peptides cyclize encoding a 29.1-kDa protein of unknown function before GS to form gramicidin S by an unknown mechanism. 1 gene, as well as 732 base pairs of an incomplete open Recently we reported the cloning and sequencing of the reading frame considered to be a part of the GS 2 gene (8). entire GS 1 gene from B. brevis Nagano (7). We have They suggested that these genes are organized in an operon. isolated four clones which expressed GS 1 activity (7). In this paper, we report the nucleotide and deduced Three of these clones, pGS309, pGS708, and pGS719 amino acid sequences of the gene behind the grs 1 gene contained about 4.6-, 8.1-, and 8.1-kb inserts, respectively, using pGS708, pGS719 and show that this gene product is immunoreactive to GS 2 antibodies and has a specific ' This work was supported in part by a research grant from the Hyogo proline-activation activity and a molecular weight of about College of Medicine. 110,000. Abbreviations: SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; IPTG, isopropyl ƒÀ-D-thiogalactoside; p-amidino PMSF, (p-amidinophenyl)methanesulfonyl fluoride.

Vol. 110, No. 1, 1991 111 112 K. Hori et al.

Sequence analysis was carried out by using the sequence

MATERIALS AND METHODS analysis software package DNASIS (Hitachi Software Engineering). Bacterial Strains and Plasmids•\Gramicidin S producer Partial Purification of the Translated Products of B. brevis Nagano and its mutant strains were used as a GS708 and GS719 in Large Scale Preparation-E. coli source of chromosomal DNA. E. coli, JM103(ƒ¢P1) JM103 cells carrying pGS708 or pGS719 were grown at [4(lac-pro); thi, strA, supE, endAsbcB, hsdR, F'traD36, 37•Ž overnight in 5 liters of L-broth supplemented with proAB, lacIq, Zƒ¢M15] or E. coli DH-1 [recA1, endA1, (100mg/ml) and IPTG (50ƒÊM). About 15g of gyrA96, thi-1, hsdR17 (rk-, mk-), supE44] were used as cells was harvested and stored at -20•Ž. Translated host cells for recombinant plasmids. products were purified from a lysate of the frozen cell paste The recombinant plasmids pGS309, pGS708, and by the procedure described previously (7). The lysate was pGS719 contained the GS 1 gene on 4.6-, 8.1-, and 8.1-kb fractionated with 50% saturated ammonium sulfate and the inserts in pUC18, respectively (7). dialyzed fraction was subjected to DEAE-Sepharose CL-6B Western Blot Analysis of the Translated Products of the column chromatography (1.5 X 45 cm) with a linear gradi Recombinant Plasmids-E. coli cells transformed by the ent of 0 to 0.28M KCl in buffer A [50mM potassium recombinant plasmids were grown at 37•Ž overnight in phosphate buffer (pH7.5), 1mM dithiothreitol, 1mM L-broth medium supplemented with ampicillin and with or MgCl2, 0.5mM EDTA, 10ƒÊM p-amidino PMSF, 10%

without IPTG as described previously (7). Total cellular glycerol]. proteins from lysed cells or the fraction from DEAE- Enzyme Assay-Activities of the amino acid-dependent Sepharose CL-6B column chromatography were subjected ATP -32PP1 exchange reaction were measured as described

to SDS-PAGE on 7.5% polyacrylamide gels as described by previously (3). Laemmli (9). Proteins were transferred to a Durapore filter Biochemicals•\[ƒ¿-32P]dCTP and 32PP1 were purchased membrane (Millipore) and the blots were treated with a from the Radiochemical Centre (Amersham, U.K.). Re 1 : 1,000 dilution of anti-GS 1 antibodies or 1:5,000 striction endonucleases were purchased from Takara Shuzo dilution anti-GS 2 antibodies and visualized with horse- (Kyoto) and Nippon Gene (Toyama). DNase and RNase radish peroxidase conjugated anti-rabbit IgG using 3,3•L-di were obtained from Sigma Chemical (St. Louis, U.S.A.). - aminobenzidine (7). The DNA and E. coli alkaline phosphatase were

Preparation of the Antibodies against Gramicidin S purchased from Takara Shuzo. All other chemicals were Synthetase 1 and Synthetase 2•\Antiserum against the standard commercial products. purified GS 1 was prepared as described previously (7). Polyclonal antibodies against the purified GS 2 were RESULTS prepared in a similar way to anti GS 1. GS 2 antibodies were further purified by using DEAE-cellulose column Detection of Immunoreactive Proteins to GS 2 Antibodies chromatography. GS 2 used for raising antibodies was in the Translated Products of the Recombinant Plasmids -Figure 1 shows the Western blot analysis of the transla purified from B. brevis as described previously (3) with some modifications. A DEAE-Sepharose CL-6B column ed products from the recombinant plasmids, pGS309, was used instead of a DEAE-cellulose column and this step pGS708, and pGS719. All these plasmids had the full was carried out before the Ornithine-Sepharose 4B column length of GS 1 gene and the sizes of the inserts were 4.6, chromatography. 8.1, and 8.1kb, respectively. DNA Sequence Analysis-DNA sequences were deter- The translated products from pGS309 included GS 1 mined for both strands by the dideoxy-chain termination protein as described previously (7). GS 1 protein was also method of Sanger et al. (10) as described previously (7). detected in the translated products from pGS708 or

Fig. 1. Western blot analysis of total cellu lar proteins from E. coli cells transformed by recombinant plasmids. Total cellular proteins present in the supernatant of lysed cells were subjected to SDS-PAGE on 7.5% poly acrylamide gels and proteins were transferred electrophoretically to Durapore membrane and reacted with GS 1 antibodies (A) or GS 2 antibodies (B), and then visualized as described in "MATERIALS AND METHODS." (A) Lane 1, lysate from E. coli JM103ĢP1 containing

pUC18; lane 2, purified GS 1 from B. brevis; lanes 3-5, lysate from E. coli JM103ĢP1 containing pGS309, pGS708, and pGS719, respectively. (B) Lane 1, lysate from E. coli JM103ĢP1 containing pUC18; lane 2, purified GS 2 from B. brevis; lanes 3-5, lysate from E. coli JM103ĢP1 containing pGS309, pGS708, and pGS719, respectively; lane 6, purified GS 1 from B. brevis; lanes 7-9, 0 to 50% (NH4)2SO4 fractions of the lysate from E. coli cells transformed by pGS708 (lane 7) and pGS719 (lanes 8 and 9). The sizes of molecular mass standards are indicated on the left. Arrows on the right show the bands of immunoreactive proteins specific to GS 2 antibodies.

J. Biochem. Proline-Activating Domain of Gramicidin S Synthetase 2 113

Fig. 2. DEAE-Sepharose CL- 6B chromatography of total cellular proteins from E. coli cells transformed by recom binant plasmids. Zero to 50% (NH4)2SO4 fraction (about 500mg of protein) of the crude extract was applied to a column (1.5 x 45cm) and eluted with a linear gradient of 0.1 to 0.28M KCl in buffer A as described in "MATERIALS AND METHODS." Fractions of 7ml each were collected. Proteins were obtained from E. coli cells carrying pGS719 (A), pGS708 (B), and pUC18 (C). - - -, protein (A280) - • - , KCl concentration; other lines show D-phenylalanine (o)-, L-proline (.)-, L-valine (A)-, L- ornithine (D)-, and L-leucine dependent ATP-32PP1 exchange activities.

Vol. 110, No. 1, 1991 114 K. Hori et al.

Fig. 3. Restriction and sequencing strategy for the grs 2 -pro gene. Restric tion fragments of the insert in clone GS708 or GS719 were subeloned into vector M13mp18 or mp19 and sequenced by the dideoxy methods (10). The se quencing strategy is indicated by arrows showing the direction and extent of DNA sequencing. The thick arrow at the top shows the direction of transcription. The hatched area indicates the coding region of GS 2.

pGS719, but less abundantly than the pGS309 product were eluted at 0.25M KCl concentration in the chroma (Fig. 1A). tographies of both pGS708 and pGS719 proteins. D-Phenyl- New translated products were detected when the blots alanine-exchange activities were not detected in the control were treated with anti GS 2 antibodies. The clones, GS708 chromatography of E. coli proteins carrying pUC18. and GS719, expressed 110-, 105-, and 115-kDa proteins, Since L-valine-, L-leucine-, and the second peak of i.e., slightly smaller proteins than GS 1 (Fig. 1B, lanes 4, 5, L-proline-dependent exchange activities were detected at 7, 8, and 9).The 110-kDa protein was more abundantly the same KCl concentrations in the control chromatogra

expressed than 115- and 105-kDa proteins. These im phy, these activities may be due to valyl-, leucyl-, and munoreactive proteins to GS 2 antibodies did not react with prolyl-tRNA synthetases, respectively. GS 1 antibodies, suggesting that these proteins were the Immunoblot analysis by dot blot revealed that im products of the grs 2 gene fragment. munoactivities to GS 2 antibodies coincided in position with The clone GS309 expressed a small amount of about 30 the first peak of L-proline-dependent exchange activities kDa protein reactive to anti GS 2 antibodies (Fig 1B, lane and immunoactivities to GS 1 antibodies coincided with 3). The molecular mass of 30 kDa agreed with the calcu D-phenylalanine-dependent exchange activities, while the lated molecular size of the GS309 product of the third open other fractions did not react with either of these antibodies reading frame which is fused to ƒÀ-galactosidase gene of (data not shown). Furthermore, the proteins of these pUC18. fractions were analyzed by means of SDS-PAGE and There were some other immunoreactive proteins to GS 2 Western blotting. The proteins of the proline-activation antibodies. One of these reacted weakly with GS 2 anti- peak reacted strongly with anti-GS 2 antibodies but did not bodies and had the same molecular mass as GS 1. Therefore react with anti-GS 1 antibodies, whereas the proteins of the it may be GS 1 detected by cross reaction of GS 2 anti- D-phenylalanine-activation peak reacted with anti-GS 1 bodies, or as a result of contamination with anti GS 1. GS antibodies and reacted weakly with anti-GS 2 antibodies 2 used for preparing GS 2 antibodies may not have been so (data not shown). The molecular masses of the immunore strictly pure as the purified GS 1 for preparation of anti GS active proteins specific to GS 2 antibodies were 110 kDa 1 antibodies. Some other proteins detected by the GS 2 (pGS708), or 115kDa and 105kDa (pGS719). These antibodies may be degradation products of the grs 2 gene molecular sizes coincided with the results from the cell product. lysate. The immunoreactive proteins to GS 1 antibodies Partial Purification of GS 2 Gene Product and Detection had the same molecular size as GS 1 from B. brevis. of Catalytic Activities Related to GS 2-Crude protein These results indicated that these immunoreactive extract was obtained by treatment from E. coli proteins to GS 1 antibodies and GS 2 antibodies are carrying pGS708 or pGS719 and fractionated with 50% products of grs 1 gene and grs 2 gene fragment, respective saturated ammonium sulfate. The dialyzed proteins were ly, and that pGS708 and pGS719 include grs 1 gene and a separated by DEAE-Sepharose CL-6B column chromatog part of grs 2 gene, the proline activation domain. raphy and each fraction was assayed for ATP-32PP1 ex DNA Sequencing Analysis-Analysis of the translated change activities dependent on the constituent amino acids product and restriction map of pGS708 and pGS719, (Fig. 2, A and B). The proteins from E. coli cells carrying indicated that GS 2 gene fragment must be located down- pUC18 were also fractionated as a control (Fig. 2C). stream from grs 1 gene. Figure 3 shows the restriction map D-Phenylalanine-, L-proline-, L-valine-, and L-leucine- and sequencing strategy for the 5•L region of grs 2 gene, grs activating activities were detected separately in the frac 2-pro gene fragment. tions of the above column chromatographies and ornithine Figure 4 shows the nucleotide sequence and deduced activation was not detected. L-Proline-dependent ATP- amino acid sequences of the grs 2 coding region overlapping 32PP , exchange activities were eluted at 0.14M KCl concen the 3•L end of grs 1 in pGS719 or pGS708 insert. There was tration and D-phenylalanine-dependent exchange activities a long open reading frame, which was separated by a

J. Biochem. Proline-Activating Domain of Gramicidin S Synthetase 2 115

Fig. 4-1

Vol. 110, No. 1, 1991 116 K. Hori et al.

Fig. 4. Nucleotide and deduced amino acid sequences of grs 2-pro gene and its 5 flanking region. The sequence data shown are for the coding region of GS 2 in the 8.1-kb insert of pGS708 or pGS719. The predicted amino acid sequence is given below the nucleotide sequence. Nucleotides and amino acids are numbered on the right and nucleotide positions are numbered from the deduced ATG initiation triplet of GS 2. A possible ribosomal is underlined. The preceding residue including the beginning of GS 2 gene (to nucleotide number 592) and the complete nucleotide sequence of GS 1 gene have already been described (7).

J. Biochem. Proline-Activating Domain of Gramicidin S Synthetase 2 117

Fig. 5. Dot matrix comparisons of grs-2 pro protein and GS 1 with GS 1 and ACVS. Check size of amino acids was 10 and matching amino acid number was 8. Comparisons were made of grs-2 pro protein with GS 1 (A), grs-2 pro protein with ACVS (B), and GS 1 with ACVS (C). Fig. 6. Comparison of the deduced primary structure (single letter code) of the proline-activation domain of GS 2 with that of GS 1 (A), and tyrocidine synthetase 2 (TS 2) (B). Colons (: ) 73-base pair noncoding sequence from grs 1 gene and between the two sequences indicate identical amino acids and dots remained open to the 3' end. A putative ribosomal binding (•) indicate chemically similar amino acids, defined as the following site was present before it. It consisted of 2,879 base pairs, groups: A, S, T, P, and G; N, D, B, E, Q, and Z; H, R, and K; M, L, coding 959 deduced amino acids. Since this frame coincided I, and V; F, Y, and W. Highly homologous regions among GS 2, GS 1, with that of the following lac Z gene, the product protein and TS 1 (A) or between GS 2 and TS 2 (B) are indicated by shading. should have been fused to lac Z protein in pGS719. Indeed, Asterisks denote the site of the conserved lysine or arginine residues. The identical amino acids in tyrocidine synthetase 1 (TS 1) are the estimated molecular weight of the fusion protein indicated by double underlines and highly conserved regions are coincided with the apparent molecular mass of 115kDa boxed in (A).The deduced amino acid sequence of grs 2-pro is that found in SDS-PAGE of the translated product (Fig. 1B, lane determined in this study, and those of GS 1 and tyrocidine '5 , 8, and 9). The 105kDa protein may be a degradation synthetases are taken from the literature (7, 8, 12). product of the 115kDa protein because the ratio of 105kDa protein to 115kDa protein increased during the purification procedure. On the other hand, as the insert was integrated agreed with the amino-terminal sequence (15 residues) of in opposite orientation to the lac Z gene in pGS708, the purified GS 2 from B. brevis except for the N-terminal open reading frame for the fusion protein terminated on the methionine (Kurotsu et al., published elsewhere). These complementary strand of lac Z gene. The estimated results suggested that this open reading frame is grs 2-pro molecular weight agreed with the apparent molecular mass gene. of 110kDa found in SDS-PAGE of the pGS708 product No consensus sequences known to bind RNA polymerase (Fig. 1B, lane 4 and 7). The deduced amino acid sequence in B. brevis are present in the intergenic noncoding region,

Vol. 110, No. 1, 1991 118 K. Hori et al.

suggesting that grs 1 and grs 2 are transcribed as one unit. the protein having ornithine-dependent ATP- 32PP1 ex Sequence Similarity of grs 2-pro Product with Grami change activity (20), and the isolation of protein fragments cidin S Synthetase 1 and Other Peptide Synthetases-A for proline-, valine-, and leucine-activating domains of GS comparison of grs 2-pro sequence with that of GS 1 2 by means of proteolytic treatments was also described revealed that there is striking similarity between the (21, 22; Kurotsu et al., published elsewhere). These C-terminal half (from 431-Cys to 959-Leu) of grs 2-pro results suggest that the activation sites of these amino acids protein and the N-terminal half (from 1-Met to 530-Phe) of consist of individual domains. GS 1, while there is no apparent similarity between the In this paper, we describe the nucleotide and deduced remaining sequences (Fig. 5A). The similarity between the amino acid sequences of a 2.9-kb fragment of grs 2-pro homologous regions was 67%, and there were some highly gene, and show that grs 2-pro gene encodes a part of GS 2, matched regions (Fig. 6A). When aligned to maximize the proline activating domain. Therefore GS 2 gene begins homology, the amino acid identities between grs 2-pro from the proline-activating domain. (residues 463 to 573, 605 to 634, 778 to 801, 806 to 883, The sequence of grs 2-pro gene had two domains. The and 928 to 957) and GS 1 (residues 33 to 146, 181 to 210, C-terminal domain was similar to the N-termimal half of 346 to 368, 373 to 450, and 495 to 524) were 84%, 80%, GS 1 and the other domain was similar to TS 2, suggesting 75%, 79%, and 86%, respectively. These regions were also that the C-terminal domain is related to common functions well conserved in tyrocidine synthetase 1 (7, 8). of these . The C-terminal half of grs-2 pro protein The C-terminal half region of grs 2-pro protein also also exhibited similarity with the three domains of the showed similarity with three regions termed domains A, B, multifunctional peptide synthetase (ACVS) from the and C of ACV synthetase (11) (Fig. 5B). When aligned to Penicillum chrysogenum (11). This enzyme catalyzes the maximize homology, the amino acid identities between grs first step of penicillin biosynthesis, the formation of the 2-pro protein (residues 400 to 959) and domains A, B, and tripeptide, ƒÂ-(L-ƒ¿-aminoadipyl)-L-cysteinyl-D-valine C of ACVS (residues 230 to 810, 1330 to 1900, and 2410 to (ACV) from L-isomers of the constituent amino acids. The 2985) were 60, 60, and 59%, respectively. acv A gene encodes a massive protein containing three The N-terminal half of GS 1 exhibited similalaritywith large, homologous domains which are similar to antibiotic each of the three conserved domains of ACVS and the peptide synthetases from Bacillus, including gramicidin S C-terminal region of GS 1 had some similarities with a synthetase 1 and tyrocidine synthetase 1. These domains region of --355 amino acids following ACVS domain C, also showed similarity with firefly luciferase (13) and which shows no similarity to grs 2-pro protein (Fig. 5C). parsley 4-coumarate-CoA ligase (14), both of which cata The sequence of grs 2-pro (residues 605-Ser to 629-Lys) lyze ATP-32PP, exchange reaction. The core sequence of was particularly well conserved in the peptide synthetases high similarity, SGTTGXPKG, is also conserved in long- [ACVS, GS 1, and tyrocidine synthetase 1 (12) ], luciferase chain acyl-CoA synthetase (15). This enzyme catalyzes (13), 4-coumarate-CoA ligase (14). A region of high simi ATP-32PP1 exchange and formation of acyl-AMP from ATP larity in this sequence, SGTTGXPKG, was also conserved and long-chain fatty acid. These results suggest that the in rat long-chain acyl-CoA synthetase (15). Other closely C-terminal region of grs-2 pro protein and homologous matched regions of grs 2-pro protein with GS 1 and regions of the other proteins contribute to activation of tyrocidine synthetase 1 were also highly conserved in amino acids or catalysis of ATP-32PP, exchange reaction. peptide synthetases (ACVS; domains A, B, and C, GS 1, Saraste et al. demonstrated a common motif in ATP- and and tyrocidine synthetase 1). It should be noted that many GTP-binding proteins, a phosphate-binding loop (P-loop) Lys or Arg residues are conserved in these highly homol- (23). SGTTGXPKG, the consensus sequence among these ogous regions of grs 2-pro, grs 1, and tyrocidine synthetase peptide synthetases may be a new class of P-loop sequence. 1 gene. These results suggest that these residues or regions On the other hand, the N-terminal half of grs-2-pro play an important role in the activities. protein had no apparent similarity with these domains of On the other hand, a comparison of grs 2-pro sequence the peptide synthetases but was highly homologous with the with the available sequence of tyrocidine synthetase 2 (TS N-terminal portion of the available sequence of tyrocidine 2) (16) showed that the N-terminal regions (from 1-Met to synthetase 2. This region may contribute to acceptance of 219-Lys) of both proteins were highly conserved. The D-phenylalanine from GS 1 or recognize proline, since GS 2 similarity was 71%. There were also some closely matched and tyrocidine synthetase 2 accept D-phenylalanine from regions (Fig. 6B). The amino acid identities between GS 1 or tyrocidine synthetase 1 and both enzymes can residues 1 to 45, 59 to 84, and 141 to 219 of these proteins activate proline, the second amino acid of the growing were 88, 78, and 75%, respectively. Since TS 2 has an pentapeptide. activity of proline activation, this region may be related to Furthermore, the C-terminal region of GS 1 and tyro specific recognition of proline. cidine synthetase 1 showed no apparent similarity with grs-2 pro protein but had some similarity with a region of DISCUSSION -355 amino acids following ACVS domain C . This region may contribute to amino acid racemization since D-valine is GS 2 activates the four constituent amino acids, that is, the third amino acid of ACV. proline, valine, ornithine, and leucine (17), but the struc The activation domains of the amino acids in the peptide ture of these activation domains has not been elucidated. synthetase genes may be aligned in the order of peptide Previously, we obtained defective GS 2, which lacked one of elongation. the specific amino acid activations, from gramicidin S Highly homologous sequences between grs 2-pro protein non-producing mutant strains of B. brevis (18, 19). Krause and GS 1 (shaded regions in Fig. 6A) were also well et al. cloned a gene fragment of GS 2 gene which expressed conserved in tyrocidine synthetase 1 and ACV synthetase.

J. Biochem. Proline-Activating Domain of Gramicidin S Synthetase 2 119

These portions involve many conserved Lys or Arg resi 10. Sanger, F., Nicklen, S., &Coulson, A.R. (1977) Proc. Natl. Acad. dues. Since arginine residues were essential for ATP Sci. U.S. 74, 5463-5467 binding and the formation of aminoacyl adenylate in GS 1, 11. Smith, D.J., Earl, A.J., & Turner, G. (1990) EMBO J. 9, 2743- 2750 GS 2, and isoleucyl tRNA synthetase (24, 25), these 12. Weckermann, R.W., FiirbaB, R., & Marahiel, M.A. (1988) conserved Lys or Arg residues may play some role in ATP Nucleic Acids Res. 16, 11841 binding or amino acid activation. 13. de Wet, J.R., Wood, K.V., DeLuca, M., Helinski, D.R., & It will be necessary to sequence the full length of the GS Subramani, S. (1987) Mol. Cell. Biol. 7, 725-737 2 gene and other multifunctional synthetase genes of 14. Lozoya, E., Hoffmann, H., Douglas, C., Schulz, W., Scheel, D., & Hahlbrock, K. (1988) Eur. J. Biochem. 176, 661-667 peptide to achieve a precise understanding of the structure and function of these enzymes. We are now 15. Suzuki, H., Kawarabayasi, Y., Kondo, J., Abe, T., Nishikawa, K., Kimura, S., Hashimoto, T., & Yamamoto, T. (1990) J. Biol. investigating the sequencing of mutant GS 1 gene lacking Chem. 265, 8681-8685 racemization activity and the full sequencing of GS 2 gene. 16. Mittenhuber, G., Weckermann, R., & Marahiel, M.A. (1989) J. Bacteriol. 171, 4881-4887 REFERENCES 17. Gevers, W., Kleinkauf, H., & Lipmann, F. (1968) Proc. Natl. Acad. Sci. U.S. 60, 269-276 1. Tomino, S., Yamada, M., Rob, H., & Kurahashi, K. (1967) 18. Iwaki, M., Shimura, K., Kanda, M., Kaji, E., & Saito, Y. (1972) Biochemistry6, 2552-2560 Biochem. Biophys. Res. Commun. 48, 113-118 2. Kanda, M., Hori, K., Kurotsu, T., Miura, S., Yamada, Y., & 19. Shimura, K., Iwaki, M., Kanda, M., Hori, K., Kaji, E., Hase Saito, Y. (1981) J. Biochem.90, 765-771 gawa, S., & Saito, Y. (1974) Biochim. Biophys. Acta 338, 577- 3. Hori, K., Kurotsu, T., Kanda, M., Miura, S., Nozoe, A., & Saito, 587 Y. (1978) J. Biochem.84, 425-434 20. Krause, M., Marahiel, M.A., von Dohren, H., & Kleinkauf, H. 4. Hori, K., Kurotsu, T., Kanda, M., Miura, S., Yamada, Y., & (1985) J. Bacteriol. 162, 1120-1125 Saito, Y. (1982) J. Biochem.91, 369-379 21. Skarpeid, H.J., Zimmer, T.L., Shen, B., & von Dohren, H. 5. Hori, K., Kanda, M., Miura, S., Yamada, Y., & Saito, Y. (1983) (1990) Eur. J. Biochem. 187, 627-633 J. Biochem.93,177-188 22. Skarpeid, H.J., Zimmer, T.L., & von Dohren, H. (1990) Eur. J. 6. Gevers, W., Kleinkauf, H., & Lipmann, F. (1969) Proc. Natl. Biochem. 189, 517-522 Acad. Sci. U.S. 63,1335-1342 23. Saraste, M., Sibbald, P.R., & Wittinghofer, A. (1990) TIBS 15, 7. Hori, K., Yamamoto, Y., Minetoki, T., Kurotsu, T., Kanda, M., 430-434 Miura, S., Okarnura, K., Furuyama, J., & Saito, Y. (1989) J. 24. Kanda, M., Hori, K., Kurotsu, T., Yamada, Y., Miura, S., & Biochem.106, 639-645 Saito, Y. (1982) J. Biochem. 91, 939-943 8. Kratzschmar, J., Krause, M., & Marahiel, M.A. (1989) J. 25. Kanda, M., Hori, K., Miura, S., Yamada, Y., & Saito, Y. (1982) Bacteriol. 171, 5422-5429 J. Biochem. 92, 1951-1957 9. Laemmli, U.K. (1970) Nature 227, 680-685

Vol. 110, No. 1, 1991