Famine Regulatory Proteins, FFRPs

2 downloads 0 Views 3MB Size Report
sus, and AsnC1 from E. coli, AsnC2 from Vibrio vulnificus, and AsnC3 from Pasturella multocidasir. Most of the eubacterial and archaeal FFRPs are separated ...
February 2008

Biol. Pharm. Bull. 31(2) 173—186 (2008)

173

Review

Transcription Regulation by Feast/Famine Regulatory Proteins, FFRPs, in Archaea and Eubacteria Tsuyoshi KAWASHIMA,a,b,c Hironori ARAMAKI,d Tomoya OYAMADA,e Kozo MAKINO,e Mitsugu YAMADA,a,b Hideyasu OKAMURA,a,b Katsushi YOKOYAMA,a,b Sanae Arakawa ISHIJIMA,a,b and Masashi SUZUKI*,a a National Institute of Advanced Industrial Science and Technology, Tsukuba Center 6–10; 1–1–1 Higashi, Tsukuba, Ibaraki 305–8566, Japan: b Japan Science and Technology Agency, Core Research for Evolutionary Science and Technology; 5 Sanbancho, Chiyoda-ku, Tokyo 102–0075, Japan: c Yokohama College of Pharmacy, Laboratory of Molecular Biology; 601 Matano, Totsuka-ku, Yokohama, Kanagawa 245–0066, Japan: d Department of Molecular Biology, Daiichi College of Pharmaceutical Sciences; 22–1 Tamagawacho, Minami-ku, Fukuoka 815–8511, Japan: and e Department of Applied Chemistry, National Defense Academy; 1–10–20 Hashirimuzu, Yokosuka, Kanagawa 239–8686, Japan. Received October 30, 2007

Feast/famine regulatory proteins (FFRPs) comprise a single group of transcription factors systematically distributed throughout archaea and eubacteria. In the eubacterial domain in Escherichia coli, autotrophic pathways are activated and heterotrophic pathways are repressed by an FFRP, the leucine-responsive regulatory protein (Lrp), in some cases in interaction with other transcription factors. By sensing the concentration of leucine, Lrp changes its association state between hexadecamers and octamers to adapt the autotrophic or heterotrophic mode. The lrp gene is regulated so that the concentration of Lrp decreases in the presence of rich nutrition. In the archaeal domain a large part of the metabolism of Pyrococcus OT3 is regulated by another FFRP, FL11. In the presence of rich nutrition, the metabolism is released from repression by FL11; transcription of fl11 is terminated by FL11 forming octamers in interaction with lysine. When the nutrient is depleted, the metabolism is arrested by a high concentration of FL11; FL11 disassembles to dimers in the absence of lysine, and repression of transcription of fl11 is relaxed. Common characteristics of the master regulations by FL11 and Lrp hint at the prototype regulation once achieved in the common ancestor of all extant organisms. Mechanisms of discrimination by FFRPs between DNA sequences and also between co-regulatory molecules, mostly amino acids, and variations of transcription regulations observed with archaea and eubacteria are reviewed. Key words

AsnC; global regulator; metabolism; phosphorylation; transcription regulation; leucine-responsive regulatory pro-

tein

INTRODUCTION The single group of transcription factors systematically distributed through archaea and eubacteria is composed of homologues of the Escherichia coli leucine-responsive regulatory protein (Lrp) (Fig. 1).1,2) Among eubacteria, E. coli has three such proteins and Pseudomonas aeruginosa (pa) eight, while, in archaea Pyrococcus sp. (P.) OT3 has fourteen such proteins, and Thermoplasma volcanium (Tv) six. Sensing the presence of rich nutrition by the concentration of leucine, E. coli shifts its metabolism from the “famine” to “feast” mode, thereby regulating 40—70 transcription units,3—5) using ca. 3000 Lrp dimers present per cell.6) In order to summarize this master regulation, Calvo and Matthews introduced the term “feast/famine regulation.”3) For this reason, Lrp and its homologues are referred to as feast/famine regulatory proteins (FFRPs).1,2) Recently, we have clarified another feast/famine regulation in the archaeal domain by an FFRP, FL11 (pot0434017) from P. OT37): see http://riodb.ibase.aist.go.jp/archaic/index. html for the “pot” and “tvg” codes used in this paper. This review firstly focuses on the molecular basis of regulation by FFRPs, before comparing regulation by FL11 with that by Lrp. In the light of the two regulations, those by other FFRPs are reviewed. It is widely accepted that the common ancestor of all extant organisms differentiated into archaea and eubacteria, with eukaryotes evolving later. From similarities ob∗ To whom correspondence should be addressed.

served between regulations by Lrp and FL11, characteristics of the prototype transcription regulation, once achieved in the common ancestor of organisms, will emerge. 3D STRUCTURE AND MOLECULAR RECOGNITION Although several FFRPs from archaea and eubacteria have been crystallized (Table 1),7—15) FL11 remains as the single FFRP whose structure has been determined in complex with its target DNA.7) A DNA duplex was included when crystallizing Lrp, but the DNA was disordered and not traced.8) Another FFRP from P. OT3, DM1 (pot1216151), was crystallized in complexes with two respective ligands that have different effects on the association state of this FFRP; isoleucine increases the association state to form octamers, and methionine has more dissociating effects to keep dimers under some conditions.9) Another FFRP, NMB0573 from the eubacterium Neisseria meningitidis has been co-crystallized with isoleucine and with methionine, although the effects of interactions have not been characterized well.10) These crystal structures have enabled us to study the molecular basis of regulation by FFRPs. Association States of FFRPs Structural units of FFRPs are dimers (Figs. 2, 3A). The N-terminal half of a full length (FL-) FFRP folds into a DNA-binding domain, DBD (Figs. 1, 3A). While, the C-terminal half folds into an assembly domain together with the identical part of the partner monomer.

e-mail: [email protected]

© 2008 Pharmaceutical Society of Japan

174

Vol. 31, No. 2

Fig. 1.

A Multiple Amino Acid Sequence Alignment of FFRPs

NMB, TTH and PF indicate NMB0573, TTH0845 and PF0864, respectively, in Table 1. Tv: Thermoplasma volcanium. pa: Pseudomonas aeruginosa. Residues at positions, where FL11 contacts DNA bases, are colored red. Basic residues found at positions 25, 39 and 41, where in FL11 the side-chains of residues bind phosphates, and Asp6 and Asp9, which in FL11 fix the orientation of Arg41 by ionic interactions, are underlined. Residues forming the Ser/Thr-Pro-Ser/Thr-Pro phase at positions 34—37 are shown on yellow backgrounds. Positions where in TvFL3 Ser/Thr residues are phosphorylated (Phos) are indicated by #. Ser/Thr residues of FFRPs found at these positions are shown on light blue backgrounds except for those in the first and third positions in the Ser/Thr-Pro-Ser/Thr-Pro phase. Asp104 and Thr/Ser at 133—135, which interact with ligand amino acid mainchains in DM1 and NMB0573, and predicted to do so in other FFRPs, are colored green on pink backgrounds. Residues at positions 80, 98, 100, 121, 122, 125 and 139, which interact with ligand amino acid side-chains or predicted to do so, are colored blue. Glu98 and Asp122 of FL11, TvFL3 and LysM, predicted to form ionic interactions with the lysine side-chain, are shown on light green backgrounds. Numbers of positions forming hydrophobic phases in a -helices are highlighted in bold.

Table 1.

3D Structures of FFRPs Determined by Crystallography

Typea) Archaeal FL FL FL

FL DM Eubacterial FL FL FL FL

DM

ID: organism

LrpAd): Pyrococcus furiosus PF0864e): Pyrococcus furiosus FL11: Pyrococcus sp. OT3

FL6: Pyrococcus sp. OT3 DM1: Pyrococcus sp. OT3

LrpC: Bacillus subtilis AsnC: Escherichia coli Lrp: Escherichia coli NMB0573: Neisseria meningtidis

TTHA0845: Thermus thermophilus

Association stateb)

DNA

Ligand

PDBc)

Ref.

24 2 (26)n 2 24* 2 24 24

None None None 17 bps None None None None

None None None None Arginine None Isoleucine Se-methionine

1I1G 2IA0 1RI7 2E1C None 2DBB 2Z4P 2E1A

11 SECSG,g) unpublished 12 7 Yamada et al. unpublished RSGI,h) unpublished 9, 15 9

24 24 24* 24 24 24 25

None None None None None None None

None Asparagine f) None None Leucine Methionine None

2CFX 2CG4 2GQQ 2P5V 2P6T 2P6S 2DJW

13 13 8 10 10 10 14

a) FL, i.e. full length having a DBD, or DM, i.e. demi, having no DBD. b) Dimeric (2), octameric (24), decameric (25) or cylinderical, with six dimers forming a helical turn, (26)n. “24*” indicates that the octamer is not closed but half open (see text). c) Protein Data Bank accession code. d) An orthologue of FL10 (pot0377090) from P. OT3. e) An orthologue of FL7 (pot0008824) from P. OT3. f) Asparagine does not affect the association state of AsnC.13) g) SECSG: South East Collaboratory for Structural Genomics. h) RSGI: Riken Structural Genomics/Proteomics Initiative.

February 2008

Fig. 2.

Various Association States of Pyrococcus FL11

Two views each of the crystal cylinder (PDB code, 1RI7) (A, B), an open octamer crystallized in the presence of arginine (C, D), and four copies of the crystal FL11 dimer in complex with DNA (PDB code, 2E1C) modeled into an octamer (E, F) in ribbon representations.

Some FFRPs such as DM1 from P. OT3 and TTHA0845 from the eubacterium Thermus thermophilus lack DBDs, corresponding to the C-terminal halves of FL-FFRPs (Fig. 1).9,14) FFRPs of this type are referred to as demi (DM-) FFRPs (Table 1).12,16) Higher-order structures are formed by assembling dimers (Fig. 2). In the crystal structures, octameric disks formed by four dimers were frequently observed (Table 1, Figs. 2E, F). In solution, intermediate states such as tetramers and hexamers are formed by DM1.9,17) A decameric disk was formed in the crystal by five dimers of TTHA0845.14) E. coli Lrp in the crystal,8) and also FL11 in the presence of arginine (Yamada M. et al., unpublished) were crystallized into octameric disks, which were not closed but half open (Figs. 2C, D). Yet in another crystal a cylinder was formed by FL11 from one end to the other end of the crystal, with six consecutive dimers forming a right-handed helical turn (Figs. 2A, B).12) Large assemblies such as ca. 20 mer or ca. 40 mer are formed in solution by some FFRPs, e.g., TvFL2 (tvg0651564) and paFL8 (gi15600501) (Shimowasa A. et al., unpublished): the “gi” code of NCBI is used for P. aeruginosa, Bacillus subtilis, and E. coli in this paper. The presence of multiple association states is one of the key characteristics important to understand the molecular basis of regulation by FFRPs. In addition to various homo-assemblies, two dimers of FL11 and two dimers of DM1 form a hetero-octamer.9) FL9

175

(pot0301583) and DM2 (pot0300646) also interact with each other (Simowasa A. et al., unpublished): genes coding the two FFRPs form an operon. Interaction of Lrp and AsnC inside the E. coli cell has been detected (Makino K. et al., unpublished). It is not known whether or not an FFRP dimer can be formed by assembling monomers of different FFRPs, or a structure equivalent to a dimer can be formed by a single polypeptide which covers two assembly domains. Mode of DNA Recognition The FL11 dimer binds to a 13 bp stretch of a DNA duplex, TGAAAWWWTTTCA, where W is T or A.7,18) Changing any base in TGAAA/ TTTCA weakens the interaction.7) Although any combinations of W bases are tolerated at the central 3 bps, replacing these by G:C severely weakens the interaction.7) In the crystal complex with TGTGAAAAATTTTCACT (Fig. 3A), a pair of DBDs in the FL11 dimer contacted two sets of 5 bps, TGAAA and TTTCA, (Figs. 3A, C) by Ala34-Thr37 (colored green in Fig. 3B) in a loop connecting a -helices 2 and 3 inside the DBD. From outside the loop, Leu24 in a -helix 2 and His39 in a -helix 3 completed hydrophobic interactions with T methyl groups in the 5 bps (Fig. 3D). The side-chains of Arg41 and His39 in a -helix 3 and Arg25 in a -helix 2 interacted with DNA phosphate groups (Figs. 3C, D). The Arg41 side-chain was directed to DNA by ionic interactions with Asp6 and Asp9; these three residues are highly conserved through FFRPs (Fig. 1). The ionic interactions and hydrogen bonds to phosphates from the Thr37 side-chain and the main-chain NH groups of Leu24 and Ala34 fixed the binding geometry in the two halves of the complex. Compared with FL11 crystallized with no DNA, the two DBDs opened relative to each other, but the opening was not enough to keep the DNA straight (Fig. 3A).7) Consequently, the DNA was bent around the FL11 dimer by ca. 55°. At the center, the DNA minor groove facing FL11 was compressed, and the major groove on the opposite side was widened. Having two hydrogen bonds only, A:T basepairs around the center produced propeller twists of ca. 20°, thereby sliding the DNA backbones relative to each other and bending the helix axis.7) Consensus binding sequences of E. coli Lrp and various other FFRPs are summarized into the form NANBNCNDNEWWWNENDNCNBNA, where e.g., NA is a base complementary to NA.19,20) This form has been fully explained by the above observations. Target DNA Sequences of FFRPs Consensus target DNA sequences have been deduced for various FFRPs (Table 2). In many cases, the numbers of examples available are small, and so they are less reliable than those of Lrp and FL11. Nevertheless, for FFRPs, where amino acid residues occupying positions 24, 34—37 and 39 are similar, e.g., for FL11 and Ss-LrpB, where residues are the same at 3 of the 6 positions, the target DNA sequences are also similar, e.g., TGAAA for FL11 and TGCTA for Ss-LrpB (Table 2). LysM, Ptr2, LrpA, Ptr1 all have Tyr/Phe at 24, Ser at 34, Glu at 35, and Arg at 39, and bind the DNA TWCGA or TWCGC.7) Partners of bases and residues able to form chemical contacts have been identified, e.g., the G base and Arg, the C base and Glu, the A base and Asn, and the T base and Ala.21) By slightly modifying the contact pattern found with the

176

Vol. 31, No. 2

Fig. 4. Contacts Made to Isoleucine and Seleno-Methionine from DM1 in the Crystal Complexes (PDB Codes, 2Z4P, 2E1A)

Fig. 3. FL11 Dimer Crystallized in Complex with the Target DNA (PDB Code, 2E1C) (A) The FL11 dimer (yellow) in complex with DNA (cyan), 5-TG[T1G2A3A4A5]A6A7T8[T9T10T11C12A13]CT-3/5-AG[T1G2A3A4A5]A6T7T8[T9T10T11C12A13]CA-3, compared with that (green) crystallized with no DNA (PDB code, 1RI7) by best overlapping Ca atoms of Met67-Ile150. The DNA in the crystal (cyan) is compared with a standard B-DNA (crimson) by best overlaying the central three basepairs. Narrowing of the minor (m) groove and widening of the major (M) groove at the center of the crystallized DNA are indicated by arrows. (B) A view looking down the DNA double helix. The main-chain of FL11 is colored green at Ala34-Thr37. (C, D) Views looking into the DNA major groove from Ala34-Thr37, in (D) showing a -helices 2 and 3. In (C) yellow arrows indicate hydrogen bonds from donors to acceptors. Red broken lines indicate hydrophobic interactions. Water molecules 1, 2, interacting with A4 and T4, and T5, respectively, form hydrogen bonds to and from Glu35. In (D) methyl groups of T3—T5 forming hydrophobic interactions with Leu24 or His39 are circled in red. Green broken lines indicate ionic interactions.

FL11–DNA complex (Figs. 3C, D), some contacts possibly formed between target DNAs and FFRPs have been predicted (Table 2)7): T at NA in DNAs and Ala or Thr at 34 and 37 in many FFRPs; G at NB and Ser/Thr at 36, and T at NB and Ala/Thr at 36; C at NC and Glu at 35, and G at NC and Arg at 35; G at ND or G complementary to C at NC and Arg at 39, C at ND or G complementary to C at NC and Gln at 39; T complementary to A at NC, ND or NE and a hydrophobic residue at 24 or 39, A complementary to T at NE and Asn at 24. Interactions with Amino Acids Many FFRPs interact with small molecules, mostly amino acids, thereby changing their association states, i.e., ligands or co-regulators of FFRPs. In the presence of leucine, Lrp changes its association state from hexadecamers to octamers.22) In the presence of lysine, FL11 assembles to octamers.7,9) Arginine facilitates hetero-octamerization by two FL11 dimers and two DM1 dimers.9) Isoleucine stabilizes homo-octamerization by DM1, while methionine destabilizes it to preserve dimers under some conditions.9,23) When DM1 was crystallized in complexes with, respectively, isoleucine and seleno-methionine, four molecules each were found inserted between a pair of tetramers (Figs. 4A,

(A, B) Top (A) and side (B) views of the DM1-isoleucine complex. Dimers related by two-fold symmetry are shown in the same colors. Interface I formed between dimers A and B (A) is divided into two halves (B), IN and IS. Into each half (e.g. IN inside a box) an isoleucine molecule (blue in IN) is inserted between a 1, b 2 and b 3 of monomer A1 and a 2 of monomer B2. The C-terminus of monomer A2 also contacts the ligand. (C, D) The isoleucine (C) and seleno-methionine (D) complexes. Colors are used for differentiating the two dimers. In (C) hydrogen bonds are indicated by broken lines with arrows indicating the donor-to-accepter directions. Two hydrogen bonds formed with Gly101 are colored red. In (D) ionic interactions are indicated by broken lines. Capital letters such as T indicate that these bonds are made with the side-chains, and lower case letters such as t135 with their main-chains. In (D) the residues where the side-chains are in hydrophobic interactions with the ligand side-chain are labeled. e.g. I139.

B). While the isoleucine side-chain was stably buried inside each pocket formed by Val98, Tyr100, Ile121, Thr122 Thr135 and Ile139 (Fig. 4C, see labels in Fig. 4D), the longer side-chain of seleno-methionine was folded (Fig. 4D). The pocket was closed by ionic interactions between Arg125 of one dimer, and Glu80 and Asp104 of another dimer (Fig. 4D). These ionic interactions will be broken, if the side-chain of seleno-methionine fully stretches. When DM1 was coinjected with methionine onto a column equilibrated with a buffer containing this amino acid, octamers of DM1 were stabilized, but DM1 was disassembled to dimers, when co-injected onto the column equilibrated with a buffer containing no methionine.9) Thus DM1 and methionine appear to be in a dynamic binding equilibrium. However, isoleucine stabilized octmerization, even when the amino acid was absent from the buffer. Although all the eight pockets formed in the crystal octamer of NMB0573 were filled with leucine or methionine,10) not all of the pockets formed in the DM1 octamer were filled with amino acids. If all the pockets were to be filled, the four dimers of DM1 would adopt a right-handed helicity and become unable to close the octamer.9) Using physico-chemical experiments, the number of lysines interacting with the FL11 octamer was calculated as ca. 4 (Shimowasa A. et al., unpublished). Another group has pointed out resemblance between the assembly domain of FFRP and the ACT domain.24,25) Although secondary structural elements composing the two

February 2008 Table 2.

177

Target DNA Sequences of FFRPs and Residues of FFRPs at the Positions Where FL11 Interacts with DNA Bases Target DNAa)

Residue in FFRP at position

FFRP

FL11 Ss-LrpB LysM Ptr2 LrpA Ptr1 FL4 LrpC Ss-Lrp AsnC Lrp AzlB PutR MdeR

Ref. NA

NB

NC

ND

NE

24

34

35

36

37

39

T T T G T/G T T/A T A T A A G T

G G T A T A T T T/A C G A C G

A§ C§ C§ C§ C§ C§ G§ C T A A T C A

A§ T G G G G A A T T A T C G

A A A A A C A A A T T A A A

Leu Leu Tyr Tyr Phe Phe Gln Met Leu Tyr Asn Asn Val Val

Ala Pro Ser Ser Ser Ser Thr Ser Ser Ser Ser Ser Ser Thr

Glu§ Ile Glu§ Glu§ Glu§ Glu§ Arg§ Pro Pro Pro Pro Pro Lys Thr

Ser Ser Ala Ser Thr Gly Gln Pro Ala Gly Thr Ser Thr Ser

Thr Thr Ala Ser Ala Thr Ala Ser Thr Thr Pro Ala Pro Pro

His Arg Arg Arg Arg Arg Ser Thr His His Leu Leu Gln Trp

7, 9 50 20 47 49, 69 47 70 20 70 20 19 20 20 20

a) NANBNCNDNE of NANBNCNDNEWWWNENDNCNBNA. Bases or their complementary bases in the same basepairs and residues contacting with each other or predicted to do so are indicated using the same expressions: e.g. for FL11, T at NA and Thr at position 37 contacting each other; G at NB and Ser at position 36; A§ at NC and ND and Glu§ at position 35; partner T bases of A at NC, ND, NE and Leu at position 24 and His at position 39.

Table 3.

Residues of FFRPs at Positions Where DM1 Interacts with the Ligand Side-Chain Residuesb) at

FFRP

Effects

a)

Ligand 98

100

121

122

135

139

80

125

104

FL11DM1 FL11 TvFL3 LysM DM2 FL9 Ss-Lrp FL4 Grp YbaO TvFL2

I I I N.C. I I N.I. I N.C. N.C. I

R KRQ K K Q Q Qc) E E Ec) C

V E E E R D D R R R I

Y T S T Y T V T A A A

L L V I H V L L Y Y V

D D D D D K K E K K L

T T T S T T T T A T T

I L L L S L L L M A F

E Y Y T E I L R L Y Q

G G R R R L K A S V R

D D D D D D D D D D D

FL5 DM3 NMB0572 TvDM DM1 Lrp paFL8 BkdR paFL3 paFL6 MdeR PutR paFL1 paFL5

I I I I I I I N.C. I N.C. N.C. N.C. I D

FILVM F, V, M, IL ML I, LFMV IVR, L, MF LA, I, V M V, I, L LM, I N.I. M P, RV ILVMI FIVML

I A A P V L L L Q L N M V L

T T T F Y S S T T S S A M T

I I V V I L L L L L T L F L

L L L I T G G D T S R G F L

T T S T T T S S S S S T S S

T V L I I M M L L I F M L L

V V R E E F F L L A A L L F

A H L R R L L T T L Q S S T

D D D D D D D D D D D D D D

a) Effects of interactions confirmed by physico-chemical experiments as increasing (I) or decreasing (D) the association state, or not confirmed (N.C.). b) Residues predicted to form ionic contacts or hydrogen bonds to the ligand side-chains are shown in the bold face, and those predicted to interact with each other between position 125 and position 80 or 104 are underlined. c) Prediction not confirmed.

units are similar, the modes of combining pairs of such elements are different: while an FFRP dimer closes a barrel, the ACT dimer is open, with the two monomers contacting each other essentially along a line. Their modes of interaction with amino acids also differ from each other. While an amino acid binds between two FFRP dimers, it binds between two ACT monomers. A Structural Code for Discriminating between Co-regulatory Amino Acids In the DM1 crystals the main-chain

OA atom of the COO terminus of isoleucine or seleno-methionine accepted hydrogen bonds from the side-chain OH groups of Thr133 and Thr135 in one DM1 dimer (Fig. 4C). Also, the side-chain OH of Ser134 was close to OB. The NH3 terminus of the ligand amino acid was contacted by the sidechain of Asp104 in another dimer by an ionic interaction (Fig. 4C). All the FFRPs known to interact with amino acids have Asp at position 104 and Thr or Ser at one of 133—135: in eubacteria Lrp interacts with leucine22,26); PutR interacts

178

with proline27); BkdR interacts with valine, isoleucine and leucine28); NMB 0572 interacts with leucine and methionine10); MdeR potentially interacts with methionine29); and Grp potentially interacts with glutamic acid30): in archaea LysM interacts with lysine31); TvFL3 (tvg1179749) also interacts with lysine9); DM2 (pot0300646) and FL9 (pot0301583) interact with glutamine9); DM3 (pot0175330), TvDM (tvg0307586) and FL5 (pot1664679) interact with hydrophobic amino acids9); FL4 (pot1613368) interacts with glutamic acid.9) Thus an FFRP having Asp104 and Thr/Ser at one of the positions 133—135 is predicted to interact with a ligand amino acid. The types of amino acids acting as ligands can be consistently explained by the types of residues occupying positions 80, 98, 100, 121, 122, 125, 135 and 139. FL11, LysM and TvFL3 all interact with lysine, and have Glu98 and Asp122 (Table 3). In DM2 Asp122 is kept but Glu98 is replaced by Arg, thereby changing the partner amino acid to glutamine. FL9 also interacts with glutamine, and has acidic and basic amino acids in the reverse order; i.e., Asp98 and Lys122. FL4 and Grp interact with glutamic acid, and have Arg98, and Lys122 or Arg80. FL5, DM3, NMB0572, DM1, TvDM, Lrp and BkdR, having many hydrophobic residues, interact with hydrophobic amino acids. Arginine also stabilizes octamerization of DM1.9) One of the residues forming the ligand-binding pocket, i.e., Ile139, is positioned at the C-terminus of the polypeptide (Fig. 4D). A model-building study has shown that the COO group of the main-chain can easily replace the hydrophobic side-chain of Ile to form ionic interactions with the two Nh atoms of arginine.9) Differences in the effects of interactions, either to increase or decrease the association state of the FFRP, are more difficult to predict, since they also depend on the basic stability of the assemblies maintained under particular conditions. However, there seems to be a pattern. Interactions with a large co-regulater, glutamic acid, stabilize octamerization of FL4, which has a small residue, Ala, at position 125 (Table 3), one of the positions closing the ligand-binding pocket in DM1. Interactions with another large amino acid, lysine, stabilize octamerization of FL11, which has the smallest amino acid Gly at position 125. However, TvFL3 and LysM have Arg125 possibly in interaction with Asp104. Interactions with lysine increase the association states of TvFL3 only to tetramers, and appear to disassemble LysM, thereby weakening its DNA-binding.31) SIMILARITIES BETWEEN REGULATIONS BY EUBACTERIAL Lrp AND ARCHAEAL FL11 We now compare global regulation of E. coli metabolism by Lrp with that of Pyrococcus metabolism by FL11. The two feast/famine regulations resemble each other in many aspects, although the regulation by Lrp can be summarized as more an instance of “heterotrophic or autotrophic,” while that by FL11 more an instance of “growth or rest.” Although Lrp interacts with other transcription factors, FL11 appears to function more independently. “Heterotrophic or Autotrophic” Regulation by E. coli Lrp E. coli has three FFRPs, Lrp, AsnC and YbaO (Table 4A). Lrp is the best characterized of all FFRPs in biological

Vol. 31, No. 2

terms. Another E. coli FFRP, AsnC, activates transcription of the asparagine synthetase gene (asnA) positioned opposite the asnC gene.32) AsnC has Ser104 instead of Asp, and its association state is not affected by asparagine.13) It is likely that the third E. coli FFRP, YbaO, is an orthologue of glutamate uptake regulatory protein (Grp),33) which activates uptake of glutamic acid in Zymomonas mobilis (see Table 3 for its potential interaction with glutamic acid).30) Lrp senses the presence of rich nutrition by interacting with leucine, and functions as a transcriptional repressor as well as an activator, thereby regulating 40–70 transcription units.3—5) Generally, biosynthetic pathways functioning in autotrophic metabolism are activated by Lrp, whereas, catabolic pathways functioning in heterotrophic metabolism are repressed. In general, repression can be achieved by blocking the approach of RNA polymerase by binding of a repressor to the promoter, or by blocking extension of the mRNA by binding of a repressor further downstream. However, for activation, an activator needs to bind upstream of the promoter to interact with RNA polymerase directly or indirectly through another factor. In the presence of leucine, the association state of Lrp changes from hexadecamers to octamers.22) This change might affect binding of Lrp to promoters: to a promoter having more than four dimer-binding sites, binding is expected to be tighter in the absence of leucine. Alternatively, interaction with leucine might cover or uncover a surface used for interaction with other transcription factors. The number of Lrp dimers present per cell changes, depending on the nutritional condition. That in E. coli cultured in a minimum medium, ca. 3000,6) decreases to ca. 2000 in a rich medium.34) The number also depends on the growth phase. At the mid-growth phase, it is decreased below 50% of that observed at the stationary phase. In short, the number increases when nutrition is depleted, thereby enhancing activation of autotrophic pathways and repression of heterotrophic pathways. In the presence of rich nutrition at the mid-growth phase, the number decreases. Consequently, heterotrophic pathways are de-repressed, while autotrophic pathways are de-activated. “Growth or Rest” Regulation by Pyrococcus FL11 Pyrococcus OT3 grows on amino acids in a hydrothermal bent, where debris of marine organisms fall from the sea above.35) Of its 14 FFRPs, 7 interact with amino acids and another FFRP possibly does so in order to regulate the metabolism in response to the amino acid availability (Fig. 5). The best characterized of the 14 FFRPs is FL11, which forms octamers in the presence of lysine (Table 5A). Generally FL11 functions as a transcriptional repressor.7) Of a set of 1025 transcription units coded in the genome, i.e., 644 independent genes and 381 operons,36) for 217 units, i.e., 21%, binding sites of the FL11 dimer were identified between TATA boxes and the 200th bases inside the first open reading frames.7) The largest population of the 217 units is involved in ATP synthesis. It has been proposed that upon degrading amino acids to carboxylic acids in Pyrococcus, ferredoxin and NAD(P)H are reduced, thereby activating NAD(P)H dehydrogenase in order to transport protons outwards in anaerobic conditions.37,38) Using the H gradient, ATP is synthesized. This entire process seems to be regulated by FL11 (Fig. 5).7)

February 2008 Table 4. A.

FFRPs from Eubacteria, E. coli (A), B. subtilis (B) and P. aeruginosa (C)

FFRPs from E. coli Abb. Lrp

Full ID

X-ray

Ligand

gi1787116 2GQQ LA, I, V interacts with AsnC in the absence of L gi1790182 2CG4 not AAs interacts with Lrp in the absence of L gi87081742 N.D. an AA an orthologue of Z. mobilis Grp

AsnC YbaO

B.

179

Association

Binding sitea)

16→8

AGAAT

ilvIH operon and other 40—70 units

8b)

TCATTc)

asnA

N.I.

N.I.

N.I.

Regulated genes

FFRPs from Bacillus subtilis Abb. bsNDM (YezC) bsFL1 (AlaR) bsFL2 (YwrC) LrpC (bsFL3)

Full ID

X-ray

Ligand

Association

Binding sitea)

Regulated genes

N.I. N.I. N.I. 8b)

N.I. N.I. N.I. TTCAW

N.I. alaRT operon N.I. lrpC

LrpA (bsFL4/YddO)

gi16077722 N.D. N.A. gi16080193 N.D. an AA gi16080664 N.D. not AAs gi16077492 2IA0 not AAs regulating sporulation and amino acid metabolism gi16077572 N.D. N.I.

N.I.

N.I.

LrpB (bsFL5/YddP)

gi16077573

N.D.

N.I.

N.I.

N.I.

AzlB (bsFL6)

gi16079725

N.D.

an AA

N.I.

AATTA

KinB dependent sporulation by regulating glyA KinB dependent sporulation by regulating glyA azlBCD operon involved in branched-chain amino acid transport

Originally identified by mutation conferring resistance to 4-azaleucine C.

FFRPs from Pseudomonas aeruginosa Abb. paFL1 paFL2 paFL3 paFL4 paFL5 paFL6 paFL7 paFL8

Full ID

Ligand

Association

Binding sitea)

Regulated genes

gi15597224 ILVMF 2→8 N.I. N.I. gi15597278 N.I. N.I. N.I. pafl2, phzM gi15597442 LM, I ca. 4→ca. 8 N.I. pafl3, phzM, phzABCDEFG an orthologue of P. putida BkdR gi15597773 not AAs N.I. N.I. phzABCDEFG, phzS gi15599160 FIVML 16→2 N.I N.I closely related with P. putida BkdR & MdeR gi15599704 an AA 2→N.I. N.I. pafl6, phzM, phzS, phzABCDEFG gi15599978 a cell component 2→N.I. N.I. phzM, phzABCDEFG, not pafl7 The ligand, a cell component, is absent when FL3/8 is over-produced, but present when FL5/7 is over-produced. gi15600501 M ca. 20→ca. 70 N.I. pafl8 E. coli Lrp orthologue

a) NANBNCNDNE of NANBNCNDNEWWWNENDNCNBNA. b) Association state observed with the crystal structure. c) A prediction but not confirmed. N.D.: not determined. N.I.: not identified. N.A.: not applicable. AA: amino acid.

Fig. 5.

Transcription Regulation of the Metabolism of P. OT3 by FFRPs

Activation and repression are indicated by arrows and T-bars respectively. Processes not fully confirmed are shown using broken lines.

The second largest population of the 217 units is involved in trans-membrane transport, followed by those involved in translation and DNA synthesis. Some other genes are involved in amino acid catabolism, e.g., aspartate oxidase (pot0038011) and malate dehydrogenase (pot0641175), in order to synthesize biochemical components using the TCA cycle. At the mid-growth phase, ca. 600 FL11 dimers are present per cell.7) This number increases to ca. 6000, 10-fold, at the stationary phase. When sensing rich nutrition as a high concentration of lysine, FL11 forms octamers, thereby terminating transcription of the fl11 gene and biosynthesis of lysine, by cooperative binding to the promoters. With the FL11 concentration decreasing, transcription of other metabolic genes is de-repressed, thereby activating catabolism of amino acids and synthesis of ATP, and shifting the metabolism into the “feast” mode. In the “famine” mode, sensing the absence of lysine, FL11 disassembles into dimers. With transcriptional repression of fl11 relaxed, ca. 6000 FL11 dimers bind to ca. 200 promoters thereby resting growth of the organism.

180

Mechanisms of Repressing and Activating Transcription by Lrp E. coli Lrp represses transcription of the lrp gene.39) The lrp promoter is positioned ca. 300 bps upstream of the first codon, and ca. 4 Lrp dimers bind around 80 to 32 by covering the 35 site. Although addition of leucine to the glucose minimum medium does not affect transcription from the lrp promoter, in a rich medium in the presence of amino acids, it is repressed by 4- to 10-fold.40) So another transcription factor appears to be involved in this regulation. Lrp activates transcription of the ilvIH operon, which encodes subunits of acetohydroxy acid synthetase, an enzyme needed for biosynthesis of isoleucine, leucine, and valine.41) In the absence of leucine, ca. 6 Lrp dimers bind upstream of the P1 promoter (D1-6 in Fig. 6B). Binding of leucine to Lrp reduces transcriptional activation by Lrp. This fact can be explained if two Lrp octamers (I and II in Fig. 6B) form a hexadecamer in the absence of leucine to recruit RNA polymerase directly or indirectly. Transcription of the papBA operon and the papI gene is regulated by Lrp (Fig. 6A).42,43) Here pap stands for pyelonephritis-associated pili. The two transcription units are positioned opposite each other, sharing a regulatory region (Fig. 6A upper). Alternative methylation at two GATC sites, 1 and 2, turns on or off transcription in both directions.

Vol. 31, No. 2

When site 2, positioned nearer the PBA promoter, is methylated, ca. 4 Lrp dimers bind nearer PI over unmethylated site 1 (oct II in Fig. 6A upper), thereby preventing approach of RNA polymerase to PI. Transcription from PBA is not activated either. When site 1 is methylated, the Lrp octamer shifts nearer to PBA (oct I) over site 2 unmethylated (Fig. 6A lower). The Lrp octamer forms an activation complex with cAMP-receptor protein (CRP), PapI, and PapB, with CRP recruiting RNA polymerase to PBA, and PapB recruiting another RNA polymerase to PI. Mechanism of Regulating Two Types of Promoters by FL11 The fl11 promoter and another promoter of the lysine-synthesis operon, coding LysJ, K, Y, Z, LeuA, B, C, D, are regulated by FL11 in response to lysine.7) In the absence of lysine two FL11 dimers bind independently to the fl11 promoter, one immediately downstream of the TATA box, and the other a further ca. 60 bps downstream inside the gene (Fig. 6C).7,9,18) In the presence of lysine, four dimers assemble into an octamer to cover ca. 110 bps in the promoter, by filling the gap between the two dimer-binding sites and by extending upstream over the TATA box. While two of the four dimers in the octamer remain to bind the specific binding-sites, the other two are expected to adopt less specific binding-modes. Similarly, in the presence of lysine, ca.

Fig. 6. Binding of Dimers of FFRPs (Blocks D0—6 and those Labeled with DM1) along Promoters, with Differentiating the Insertions between Blocks to 7—8 bps (Shorter Bars) and ca. 18 bps (Longer Bars) (A) Lrp on the pap promoter, where GATC site 1 or 2 is methylated alternatively. Transcription of papBA and papI is activated (lower), when GATC site 1 is methylated and the Lrp octamer occupies the “oct I” position, but not activated, when GATC site 2 is methylated and the Lrp octamer occupies the “oct II” position (upper). (B) Lrp on the ilvIH promoter. Transcription is activated in the absence of leucine, when at least six Lrp dimers bind to the promoter, possibly by forming a hexadecamer by interaction of octamers I and II (see text). Binding of D0 and D2B are not detected by foot-printing experiments,71) but these binding sites might be present.2) (C) FL11 alone or FL11 plus DM1 on the fl11 promoter. In the absence of lysine, two FL11 dimers bind. In the presence of lysine, an FL11 octamer, formed by four dimers, binds and terminates transcription. In the presence of arginine two FL11 dimers and two DM1 dimers bind. (D) Ptr2 on the fdxA promoter to activate transcription of fdxA and to repress that of ptr2.

February 2008

110 bps in the lysine-synthesis promoter are covered.7,9) Transcription from the two promoters is terminated in vitro in the presence of lysine, while in its absence, FL11 octamers disassemble to dimers, and transcriptional repression by FL11 is relaxed.7) Other simpler metabolic promoters do not have octamerbinding sites. In the presence of rich nutrition, transcription of fl11 is terminated, and thus the number of FL11 octamers present per cell decreases. Simpler promoters are unable to compete with the fl11 and lysine-synthesis promoters for interaction with FL11. Consequently, transcription from metabolic promoters is activated. When the nutrient is depleted, transcriptional repression of fl11 is relaxed. An increasing number of FL11 dimers arrest growth of the organism. When lysine is absent but arginine is present, in order to prevent inappropriate shifting to the famine mode, the FL11DM1 hetero-octamer is formed (Fig. 6C), thereby keeping transcription of the fl11 gene repressed. In the lysine-synthesis promoter, most of the block is positioned upstream of the TATA box except for an FL11 dimer-binding site positioned overlapping the TATA box. Although FL11-DM1 also binds to the lysine-synthesis promoter in the presence of arginine, from around a site facing one of the two DM1 dimers, transcription factor B and the TATA-binding protein (TBP) might enter to activate biosynthesis of lysine.9) Similarly to FL11, some other archaeal FFRPs bind on top or downstream of TATA boxes, possibly to repress transcription, e.g., Ss-Lrp on the ss-lrp operator in Sulfolobus solfataricus,44) and LrpA on the lrpA operator in Pyrococcus furiosus.45,46) Mechanism of Activating Transcription by Ptr2/LrpA/ FL10 While FL11 functions as a transcriptional repressor, FL10 (pot0377090) appears to be functional as a transcriptional activator. This FFRP is an orthologue of LrpA from P. furiosus. Furthermore, it is likely that the two FFRPs might be orthologues of Ptr2 (Suzuki M. unpublished). Ptr2 is an FFRP from Methanococcus jannaschii.47) In vitro transcription from the fdxA and rb2 promoters is activated by Ptr2.48) A pair of Ptr2 dimers bind to each promoter, with the nearest dimer positioned upstream of the TATA box with an insertion of 6—8 bps. The fdxA gene is positioned opposite the ptr2 gene, and the two units share an intergenic region (Fig. 6D). The fdxA promoter is positioned inside the ptr2 gene, so that activation of the fdxA necessarily represses transcription of ptr2. Similarly to Ptr2, two dimers of FL10/LrpA bind downstream of the TATA box in the promoter of fl10/lrpA.45,46) No homologue of rb2 is coded in the genome of P. OT3 but another homologue, the rubredoxin gene (pf1282), is present in the genome of a closely related archaeon, P. furiosus. By genomic SELEX experiments, the promoter of pf1282 was selected as bound by FL10/LrpA.49) Two homologues of fdxA are present in the genome of P. OT3, which encode the d subunits of 2-ketoisovalerate oxidoreductase (VOR) (pot1184199) and pyruvate oxidoreductase (POR) (pot1181715). The two genes are part of a single operon, encoding subunits of VOR and POR, which reduce ferredoxin, but are not directly regulated by FL11 (Fig. 5). Binding sites of FL10 are positioned upstream of TATA boxes in the vor/por (Yokoyama K. et al., unpublished) and pf1282 promoters.48) Since the fl10 gene is regulated by FL11

181

(Yokoyama K. et al., unpublished), synthesis of VOR and POR is indirectly regulated by FL11, so that it is activated in the presence of lysine (Fig. 5). Some other archaeal FFRPs also bind upstream of TATA boxes, e.g., Ss-LrpB to the ss-lrpB promoter,50) and LysM to the lysM promoter,51) in Sulfolobus solfataricus. These FFRPs might also function as transcriptional activators. Hierarchies of Regulations by FFRPs in Pyrococcus FL11 from Pyrococcus and TvFL3 from Thermoplasma in the subdomain Euryarchaeota, and LysM from Sulfolobus in the other subdomain Crenarchaeota, all interact with lysine. It is difficult to explain why response to availability of lysine is so important for archaea. In contrast, no eubacterial FFRP has been identified as interacting with lysine. For Pyrococcus, availability of glutamic acid or glutamine appears to be more important than that of lysine. The first step of synthesizing ATP is to degrade amino acids to a -keto acids. For this reaction 2-oxoglutarate is needed, which can be synthesized from glutamic acid by glutamate dehydrogenase (GDH, pot0376559). Glutamic acid can be synthesized from glutamine by glutamine synthetase (pot1467407). The same reaction can be catalyzed by asparagine synthetase (pot0787200), while converting aspartic acid to asparagine. So far no FFRP from P. OT3 has been identified as interacting with asparagine or aspartic acid, but FL4 interacts with glutamic acid, and FL9 and DM2 interact with glutamine.9) Also FL11 weakly interacts with glutamine.7,9) In addition to fl10, three other genes coding FFRPs, fl6 (pot1735659), fl7 (pot0008824) and fl8 (pot0123002), have FL11 dimer-binding sites in their regulatory regions. It is likely that regulation by FFRPs is organized in some hierarchies, depending on the metabolic importance of amino acids they interact with (Fig. 5). Possible Roles of the FL11 Cylinder In one of the crystals of FL11 a cylinder was formed by spanning ca. 2 mm from one end to the other end of the crystal.12) Although formation of such a cylinder has not been observed in vivo so far, it might shut down the entire metabolism in an emergency, e.g., when the cell is exposed to a temperature too high or too low. When the P. OT3 genome is represented by a clock diagram,52) the density of 13 bp sites, having 0—3 bp mismatches on TGAAAWWWTTTCA, is highest at ca. 7:00 with flanking regions at ca. 5:00 and ca. 9:30 (Suzuki M. et al., unpublished). The density of transcription and translation-related genes is highest at ca. 1:30, while, that of ATP synthesis-related genes regulated by FL11 is highest at ca. 3:30 near the fl11 gene positioned at ca. 2:30. If cylinder formation is initiated at ca. 7:00 and propagates through the two halves to ca. 1:30, it will successively terminate ATP synthesis, transcription of fl11, and finally transcription and translation of the whole set of genes. Formation of a cylinder might not be unique to FL11. When Lrp was crystallized, its octameric disk was halfopen.8) The geometry between Lrp dimers is an intermediate between that in a closed octamer and that in the Fl11 cylinder. Large assemblies such as ca. 20 mer or ca. 40 mer are formed in solution by some other FFRPs, e.g., TvFL2 and paFL8 (Tables 4, 5). Furthermore, in Bacillus subtilis some FFRPs are involved in the process of sporulation (see next section), which can be regarded as another preparation for an emergency.

182

Vol. 31, No. 2

Table 5. A.

FFRPs from Archaea, P. OT3 (A) and T. volcanium (B)

FFRPs from Pyrococcus sp. OT3 Abb. DM1

DM2 DM3 FL1 FL2 FL3 FL4 FL5 FL6 FL7 FL8 FL9 FL10 FL11

B.

Full ID

X-ray

Ligand

Association

pot1216151

2E1A, IVR, L, MF 2→8 2Z4P interacting with FL11 in the presence of arginine for repressing fl11 pot0300646 N.D. Q N.I. interacting with FL9, with dm2 and fl9 forming an operon pot0175330 N.D. F, V, M, IL 2→8 pot0828564 N.D. not AAs N.I. not coded in the genome of P. abyssi or P. furisosus pot0836696 N.D. N.I. N.I. not coded in the genome of P. abyssi or P. furisosus pot0868477 2IA0 not AAs 2b) an orthologue of P. furiosus PF0864 pot1613368 N.D. E 2→8 pot1664679 N.D. FILVM 4→8 pot1735659 2DBB not AAs 2b) pot0008824 N.D. N.I. N.I. pot0123002 N.D. an AA N.I. pot0301583 N.D. Q N.I. interacting with DM2, with dm2 and fl9 forming an operon pot0377090 1I1G not AAs 8b) an orthologue of P. furiosus LrpA and M. jamnaschii Ptr2 pot0434017 2E1C, KRQ 2b)→8 1RI7 cylinderb) interacting with DM1 in the presence of arginine to repress fl11

Binding sitea)

Regulated genes

N.A.

fl11

N.A.

N.I.

N.A. N.I.

N.I. N.I.

N.I.

N.I.

N.I.

N.I.

A(T)TGAA N.I. N.I. N.I. N.I. N.I.

fl4, flavoprotein (pot0803503), signal recognition particle (pot0175333) N.I. N.I. N.I. N.I. N.I.

T(G)TCGA

fl10, rubredoxin (pf1282)

TGAAA

fl11, fl10, lysine-synthesis operon, h-atpase, nad(p)h dehydrogenase etc.

FFRPs from Thermoplasma volcanium Abb. TvDM TvFL1 TvFL2 TvFL3 TvFL4 TvFL5

Full ID

Ligand

tvg0307586 I, LFMV synthesized constitutively tvg0368274 O2 synthesized in aerobic conditions tvg0651564 an AA synthesized more in aerobic conditions tvg1179749 K phosphorylated at 8 Ser/Thr tvg1254131 not AAs tvg1409444 N.I. synthesized more in anaerobic conditions

a) NANBNCNDNE of NANBNCNDNEWWWNENDNCNBNA. plicable. AA: amino acid.

Association

Binding sitea)

Regulated genes

2→8

N.A.

N.I.

2→8

N.I.

porphobilinogen synthase (tvg1155064)

4—6, ca. 20

N.I.

2→4

N.I.

4—8, 10—14, 20—40 2, 4—6, 8—10

N.I.

Rieske ion sulfur proteins (tvg0360023, tvg0366346), nadh dehydrogenase Rieske ion sulfur proteins (tvg0360023, tvg0366346), nadh dehydrogenase N.I.

N.I.

Rieske ion sulfur proteins (tvg0360023, tvg0366346), nadh dehydrogenase

b) Association state observed with the crystal structure. N.D.: not determined. N.I.: not identified. N.A.: not ap-

REGULATION BY VARIOUS OTHER FFRPS IN EUBACTERIA AND ARCHAEA In many archaea and eubacteria, FFRPs regulate biosynthesis or catabolism of amino acids through transcription regulation of genes coding dehydrogenases, i.e., oxide reductases such as VOR. Hydrophobic amino acids such as leucine and methionine, and polar residues such as lysine, glutamic acid and glutamine function as co-regulators of FFRPs. Regulation by Bacillus FFRPs Bacillus subtilis has six FL-FFRPs (bsFL1—6) and an N-demi FFRP (bsNDM), which corresponds to the N-terminal DBD of an FL-FFRP (Table 4B).53) Similar NDM-FFRPs are found with archaea such as Sulfolobus solfataricus. bsNDM has Asp6, Asp9 and Arg41, which are important in fixing the DNA-binding geometry of FL11. Sporulation of B. subtilis is regulated by two membranebound kinases, KinA and B. bsLrpA (bsFL4) and bsLrpB

(bsFL5) are involved in the processes regulated by KinB, by regulating transcription of the glyA gene.54) The glyA gene encodes serine hydroxymethyl transferase, which catalyzes interconversion between serine and glycine. bsLrpC (bsFL3) regulates bslrpC gene, and it is possibly involved also in sporulation.55) AzlB (bsFL6) regulates the azlBCD operon, which is involved in transport of branchedchain amino acids, isoleucine, valine and leucine.56) Regulation by Pseudomonas FFRPs The eubacterium Pseudomonas aeruginosa has eight FFRPs (Table 4C), and this number is the largest among eubacteria.16,57) paFL3 (gi15597442) is an orthologue of BkdR from Pseudomonas putida (Table 6).28) P. putida has pathways for oxidizing many amino acids.58) BkdR regulates the bkd operon, coding four polypeptides of a -ketoacid dehydrogenase, or oxide reductase (Table 6), involved in catabolism of isoleucine, valine and leucine.28) paFL5 (gi15599160) also is closely related with BkdR, and another FFRP, MdeR. MdeR from P.

February 2008 Table 6.

183

Miscellaneous FFRPs from Archaea and Eubacteria

ID LysM Ss-Lrp

Organism

A/Ea)

FL/DM

X-ray

Ligand

Ass.b)

Binding sitec)

Regulated genes

N.D. N.D.

K an AA, e.g., Q

N.I. 8

TTCGA ATCTT

lysine-synthesis ss-lrp

N.D. N.D. N.D.

N.I. N.I. not AAs

N.I. N.I. N.I.

TGCTA TACGC GACGA

ss-lrpB N.I. ptr2, fdxA, rb2

Grp BkdR

Sulfolobus solfataricus A FL Sulfolobus solfataricus A FL An orthologue of Sa-Lrp from Sulfolobus acidocaladarius Sulfolobus solfataricus A FL Methanococcus jannaschii A FL Methanococcus jannaschii A FL An orthologue of P. furiosus LrpA and P. OT3 FL10 Zymomonas mobilis E FL Pseudomonas putida E FL

N.D. N.D.

E V, I, L

N.I. N.I.

N.I. N.I.

MdeR

Pseudomonas putida

E

FL

N.D.

M

N.I.

TGAGA

PutR

E

FL

N.D.

P, RV

N.I.

GCCCA

NMB 0573

Rhodobacter capulatus Agrobacterium tumegaciens Neisseria meningtidis

E

FL

2P5V, 6S, 6T

ML

ca. 8→8

N.I.

TTH 0845

Thermus thermophilus

E

DM

2DJW

not AAs

N.I.

N.A.

gltP, grp bkd operon catabolizing V, I, L mdeAB operon catabolizing M proline dehydrogenase (putA) ftsZ, murI, porA, B, pileO, P, opc, opa, mafB N.I.

Ss-LrpB Ptr1 Ptr2

a) Archaeal (A) or eubacterial (E). b) Association state. c) NANBNCNDNE of NANBNCNDNEWWWNENDNCNBNA. N.D.: not determined. N.I.: not identified. N.A.: not applicable. AA: amino acid.

putida regulates the mdeAB operon.29) The mdeA gene encodes L-methionin-g -lyase, while the mdeB gene encodes the E1 component of a dehydrogenase, possibly involved in catabolism of methionine.29) paFL8 is closely related to E. coli Lrp.57) Although amino acid residues occupying the positions predicted to interact with its co-regulatory amino acid are almost the same as those in E. coli Lrp (Table 3), at the eight positions predicted to interact with DNA bases, paFL8 is not the same as Lrp. Instead, paFL6 (gi15599704) closely resembles Lrp (Fig. 1). The association states of paFL1 (gi15597224), paFL3, paFL5 and paFL8 are affected by the presence of hydrophobic amino acids (Table 3, Ishijima S. A. et al., unpublished). In addition, paFL6 is predicted to interact with an amino acid. When genes coding FFRPs were introduced into P. aeruginosa using a plasmid, the color of the media changed from yellow-greenish when culturing the standard strain PAO1 to blue-greenish, when culturing a strain overproducing paFL8, 5, 7, or 2 or more yellowish, when culturing a strain overproducing paFL6, 4, 3 or 1 (Aramaki H. et al., unpublished). These observations suggest that FFRPs 5, and 8 might regulate biosynthesis of dyes, possibly from amino acids. When a medium used to culture the standard strain PAO1 was added to a solution of paFL7 (gi15599978), the association state of paFL7 increased. Such a change was not observed when a fresh medium was added to the solution. An unidentified ligand appears to be released from the standard strain. This ligand appears to be present in a medium used for culturing a strain overproducing paFL7 or 5, but not present in that culturing a strain overproducing paFL3 or 8 (Shimowasa A. et al., unpublished). Regulation by Thermoplasma FFRPs The arcaheon Thermoplasma volcanium can survive in aerobic as well as anaerobic conditions.59) In aerobic conditions, thioredoxin peroxidase (tvg0230902), hydroperoxide reductase (tvg1154375) and superoxide dismutase (tvg0062871) are synthesized to reduce active oxygens (Fig. 7).60) Enzymes, 2-oxoglutarate oxidoreductase (KGOR) (tvg1406112, 1405258), pyruvate dehydrogenase E1 component b subunit (tvg-

Fig. 7. Proteins Synthesized in Aerobic (Left) and Anaerobic (Right) Conditions, Separated by pI Values (Horizontal) and MWs (Vertical), Using 2D Gel Electrophoresis Hydroperoxide reductase (tvg1154375) (1) and thioredoxin peroxidase (tvg0230902) (2) are synthesized more in an aerobic condition, while 2-oxoglutarate oxidoreductase (KGOR) b subunit (tvg1405258) (4) and pyruvate dehydrogenase E1 component b subunit (tvg0105141) (3) more in an anaerobic condition.

0105141), electron transfer flavoprotein (tvg1345079, 1344033) and NADH dehydrogenase C subunit (tvg1165354), are synthesized in anaerobic conditions (Fig. 7).60) This observation can be explained if, in anaerobic conditions, ATP is synthesized by reactions similar to those of Pyrococcus. While, Rieske iron sulfur binding proteins (tvg0360023, 0366346) appear to be the key enzymes for producing a Hgradient in aerobic conditions. Of 6 FFRPs present in T. volcanium (Table 5B), TvFL1 (tvg0368274) appears to be important to the aero/anaero adaptation (Kawashima T. et al., unpublished). It is synthesized exclusively in aerobic conditions. TvFL1 forms octamers in the presence of oxygen molecules, while in their absence it disassembles to dimers. His and Cys residues are present in the assembly domain around positions likely to in-

184

teract with its co-regulator, suggesting that TvFL1 might interact with oxygen molecules, through metals such as iron bound by the His/Cys residues. TvFL1 binds to the promoter of porphobilinogen synthetase (tvg1155064) (Kawashima et al., unpublished). TvFL2 is produced more in aerobic conditions, and TvFL5 (tvg1409444) more in anaerobic conditions, while TvFL3 to similar degrees in the two types of environments (Kawashima T., unpublished). These three FFRPs bind to the promoters of Rieske ion sulfur proteins and nadh dehydrogenase. TvFL3 forms tetramers in the presence of lysine,9) while TvFL2 forms ca. 14 mer in the presence of 20 amino acids, although its partner amino acid has not been identified (Shimowasa A. et al., unpublished). TvDM is synthesized in both types of environments,60) and forms octamers in the presence of a hydrophobic amino acid such as isoleucine.9) Regulations by Other Miscellaneous FFRPs An FFRP, PutR, distributes through a -proteobacter such as Rhodobacter and Agrobacterium.61,62) It regulates the putA gene encoding proline dehydrogenase. AldR from the leguminous bacterium Rhizobium leguminosarium, regulates the aldA gene encoding L-alanine dehydrogenase. Inside the bacteroid in symbiosis with pea, the enzyme is used possibly in order to optimize the concentration of alanine.63) When the bacterium is in free-living culture, AldA is induced best by alanine, but also by carboxylic acids such as succinate, malate and pyruvate.63) AldR has Asp at position 104, and Ser at position 135. Of the two FFRPs present in the archaeon M. jannaschii, Ptr2 regulates rb2 and fdxA.49) As has been described, fdxA encodes a homologue of d subunits of VOR and POR. Rubredoxin is another non-heme iron-binding protein, and similarly to ferredoxin, it has an oxide reductase activity. The amino acid sequence of the other FFRP, Ptr1,48) resembles that of Ptr2. Another archaeon, S. solfataricus, has seven FL-FFRPs and two DM-FFRPs.16) Of the FL-FFRPs, Ss-Lrp and SsLrpB regulate transcription of genes encoding these proteins, respectively. Ss-Lrp has Asp at position 98 and Lys at position 122, and it is predicted to interact with glutamine (Table 3). LysM regulates an operon coding enzymes synthesizing lysine, LysJ, K, W, X, Y, Z, and also LysM.31) In addition, 6 NDM-FFRPs are coded in the genome.53) One of the 6 NDM-FFRPs is Lrs14, which regulates the lrs14 gene.64) These NDMs have Asp at position 9 but Ser at position 6. Instead of Arg many of them have Ser at position 41. SUMMARY: FROM THE LAST COMMON ANCESTOR TO EUKARYOTES From the Last Common Ancestor to Archaea and Eubacteria On the basis of what has been discussed in this paper, it is likely that the ancestor FFRP (or FFRPs) regulated metabolism of the last common ancestor of all extant organisms, in particular, catabolism and biosynthesis of biochemical components, through genes encoding dehydrogenases, possibly in interaction with a small molecule such as an amino acid or its derivative. When the last common ancestor differentiated to eubacteria and archaea, RNA polymerase differentiated to what is

Vol. 31, No. 2

Fig. 8. FFRPs

Phylogeny of Archaeal (Underlined) and Eubacterial (Italic)

Orthologues cluster with each other, e.g., FL10 from P. OT3 and LrpA from P. furiosus, and AsnC1 from E. coli, AsnC2 from Vibrio vulnificus, and AsnC3 from Pasturella multocidasir. Most of the eubacterial and archaeal FFRPs are separated from each other into two clusters (indicated as “Eubacterial” and “Archaeal”). Notable exceptions are AsnC 1—3 of the eubacterial origin clustering with archaeal FFRPs.

now specific to eubacteria and what is now shared by archaea and eukaryotes. Associating factors differentiated to sigma factors and other activators in eubacteria and TBP and TFB shared in archaea and eukaryotes. FFRPs in the two domains have differentiated so that they interact with these different proteins. Differentiation of FFRPs can be analyzed by identifying each orthologous group by a cluster analysis of amino acid sequences, so that inside each group the topological relation of the sequences coincides with phylogenic relations between the organisms.65) In this way, it has been concluded that Lrp is not distributed outside b - and g -proteobacter, and is closely related with another FFRP distributed through a -proteobacter, i.e., a PBFs. While, AsnC is most different from other eubacterial FFRPs. In fact, AsnC resembles archaeal FFRPs such as FL11 and FL10/LrpA (Fig. 8). For this reason, archaeal FFRPs were once referred to as AsnC rather than Lrp.66) It is likely that a prototype FFRP has differentiated to AsnC and another, and the latter differentiated to YbaO/Grp and Lrp/a PBF.65) From Archaea and Eubacteria to Eukaryotes If eukaryotes have been created by endo-symbiosis of archaea and eubacteria, since both archaea and eubacteria have FFRPs, variants of FFRPs might be still functioning in some eukaryotes. FFRPs comprise a super-assembly of many orthologous groups, and in general the homology between two paralogous FFRPs is less than 30%. Without knowing the 3D structure, it is difficult to identify a eukaryotic FFRP solely on the basis of its amino acid sequence. Many transcription factors have DBDs composed of three a -helices, and FFRPs are not exceptions. In terms of the

February 2008

geometry of combining sets of three a -helices in DBDs, homeodomains of transcription factors regulating body formation of multi-cellular eukaryotes closely resemble DBDs of FFRPs. In a sense, master regulation by transcription factors having homeodomains is comparable with another master regulation by FL11 (Fig. 5). It can be important to know whether or not this resemblance reflects any evolutionary processes. Many proteins extracted from the archaeon T. volcanium are phosphorylated.60) Eight Ser/Thr residues have been identified as phosphorylated in TvFL3,7) all inside the DBD (marked # in Fig. 1, Phos). Such phosphorylation will weaken interaction with DNA. Ser34 and Thr37 are positioned inside the DNA-recognition loop. This Thr37, present also in FL11, and another phosphorylation site, Thr25, replaced by Arg in FL11, contacted DNA phosphates in the crystal.7) In many FFRPs, positions 34—37 in their DNA recognition loops have tendencies to form a Ser/Thr-Pro-Ser/Thr-Pro phase (shown on the orange background in Fig. 1). This phasing resembles the Ser-Pro-X-X motif found in many eukaryotic transcription factors at their phosphorylation sites.67,68) Thus the origin of eukaryotic modulations of transcription is traced back to archaea. Acknowledgements This work was supported by Core Research for Evolutionary Science and Technology program of Japan Science and Technology Agency. REFERENCES 1) Yokoyama K., Ishijima S. A., Clowney L., Koike H., Aramaki H., Tanaka C., Makino K., Suzuki M., FEMS Microbiol. Rev., 30, 89— 108 (2006). 2) Suzuki M., Proc. Jpn. Acad., 79B, 274—289 (2003). 3) Calvo J. M., Matthews R. G., Microbiol. Rev., 58, 466—490 (1994). 4) Wagner R., “Transcription Regulation in Prokaryotes,” Oxford University Press, Oxford, 2000. 5) Newman E. B., Lin R., Annu. Rev. Microbiol., 49, 747—775 (1995). 6) Willins D. A., Ryan C. W., Platko J. V., Calvo J. M., J. Biol. Chem., 266, 10768—10774 (1991). 7) Yokoyama K., Ishijima S. A., Koike H., Kurihara C., Shimowasa A., Kabasawa M., Kawashima T., Suzuki M., Structure, 15, 1542—1554 (2007). 8) de los Rios S., Perona J. J., J. Mol. Biol., 366, 1589—1602 (2007). 9) Okamura H., Yokoyama K., Koike H., Yamada M., Shimowasa A., Kabasawa M., Kawashima T., Suzuki M., Structure, 15, 1325—1338 (2007). 10) Ren J., Sainsbury S., Combs S. E., Capper R. G., Jordan P. W., Berrow N. S., Stammers D. K., Saunders N. J., Owens R. J., J. Biol. Chem., 282, 14655—14664. 11) Leonard P. M., Smits S. H. J., Sedelnikova S. E., Brinkman A. B., de Vos W. M., van der Oost J., Rice D. W., Rafferty J. B., EMBO J., 20, 990—997 (2001). 12) Koike H., Ishijima S. A., Clowney L., Suzuki M., Proc. Natl. Acad. Sci. U.S.A., 101, 2840—2845 (2004). 13) Thaw P., Sedelnikova S. E., Muranova T., Wiese S., Ayora S., Alonso J. C., Brinkman A. B., Akerboom J., van der Oost J., Rafferty J. B., Nucl. Acids Res., 34, 1439—1449 (2006). 14) Nakano N., Okazaki N., Satoh S., Takio K., Kuramitsu S., Shinkai A., Yokoyama S., Acta Crystallogr., F62, 855—860 (2006). 15) Kudo N., Allen M. D., Koike H., Katsuya, Y., Suzuki, M., Acta Crystallogr., D57, 469—471 (2001). 16) Koike H., Sakuma M., Mikami A., Amano N., Suzuki M., Proc. Jpn. Acad., 79B, 63—69 (2003). 17) Sakuma M., Nakamura M., Koike H., Suzuki M. Proc. Jpn. Acad., 81B, 110—116 (2005).

185 18) Yokoyama K., Ebihara S., Kikuchi T., Suzuki M., Proc. Jpn. Acad., 81B, 64—75 (2005). 19) Cui Y., Wang Q., Stormo G. D., Calvo J. M., J. Bacteriol., 177, 4872— 4880 (1995). 20) Suzuki M., Proc. Jpn. Acad., 79B, 213—222 (2003). 21) Suzuki M., Structure, 3, 317—326 (1994). 22) Chen S., Calvo J. M., J. Mol. Biol., 318, 1031—1042 (2002). 23) Sakuma M., Koike H., Suzuki M., Proc. Jpn. Acad., 81B, 26—32 (2005). 24) Ettema T. J. G., Brinkman A. B., Tani T. H., Rafferty J. B., van der Oost J., J. Biol. Chem., 277, 37464—37468 (2002). 25) Brinkman A. B., Ettema T. J. G., de Vos W. M., van der Oost J., Mol. Microbiol., 48, 287—294 (2003). 26) Marasco R., Varcamonti M., La Cara F., Ricca E., de Felice M., Sacco M., J. Bacteriol., 176, 5197—5201 (1994). 27) Jafri S., Evoy S., Cho K., Craighead H. G., Winans S. C., J. Mol. Biol., 288, 811—824 (1999). 28) Madhusudhan K. T., Luo J., Sokatch J. R., J. Bacteriol., 181, 2889— 2894 (1999). 29) Inoue H., Inagaki K., Eriguchi S., Tamura T., Esaki N., Soda K., Tanaka H., J. Bacteriol., 179, 3956—3962 (1997). 30) Peekhaus N., Tolner B., Poolman B., Krämer R., J. Bacteriol., 177, 5140—5147 (1995). 31) Brinkman A. B., Bell S. D., Lebbink R. J., de Vos W. M., van der Oost J., J. Biol. Chem., 277, 29537—29549 (2002). 32) Kölling R., Lother H., J. Bacteriol., 164, 310—315 (1985). 33) Bédu S., Joset F., Mol. Microbiol., 24, 230 (1997). 34) Landgraf J. R., Wu J., Calvo J. M., J. Bacteriol., 178, 6930—6936 (1996). 35) González J. M., Masuchi Y., Robb F. T., Ammerman J. W., Maeder D. L., Yanagibayashi M., Tamaoka J., Kato C., Extremophiles, 2, 123— 130 (1998). 36) Amano N., Tsuji K., Suzuki M., Proc. Jpn. Acad., 79B, 131—136 (2003). 37) Sapra R., Bagramyan K., Adams M. W. W., Proc. Natl. Acad. Sci. U.S.A., 100, 7545—7550 (2003). 38) Adams M. W. W., Holden J. F., Menon A. L., Schut G. J., Grunden A. M., Hou C., Hutchins A. M., Jenny F. E., Jr., Kim C., Ma K., Pan G., Roy R., Sapra R., Story S. V., Verhagen M. F. J. M., J. Bacteriol., 183, 716—724 (2001). 39) Wang Q., Wu J., Friedberg D., Plakto J., Calvo J. M., J. Bacteriol., 176, 1831—1839 (1994). 40) Lin R., D’Ari R., Newman E. B., J. Bacteriol., 174, 1948—1955 (1992). 41) Platko J. V., Willins D. A., Calvo J. M., J. Bacteriol., 172, 4563—4570 (1990). 42) Braaten B. A., Plakto J. V., van der Woude M. W., Simons B. H., de Graaf F. K., Calvo J. M., Low D. A., Proc. Natl. Acad. Sci. U.S.A., 89, 4250—4254 (1992). 43) Nou X., Braaten B., Kaltenbach L., Low D. A., EMBO J., 14, 5785— 5797 (1995). 44) Enoru-Eta J., Gigot D., Glansdorff N., Charlier D., Mol. Microbiol., 45, 1541—1555 (2002). 45) Brinkman A. B., Dahlke I., Tuininga J. E., Lammers T., Dumay V., de Heus E., Lebbink J. H. G., Thomm M., de Vos W. M., van der Oost, J., J. Biol. Chem., 275, 38160—38169 (2000). 46) Dahlke I., Thomm M., Nucl. Acids Res., 30, 701—710 (2002). 47) Ouhammouch M., Geiduschek E. P., EMBO J., 14, 5785—5797 (2001). 48) Ouhammouch M., Dewhurst R. E., Hausner W., Thomm M., Geiduschek E. P., Proc. Natl. Acad. Sci. U.S.A., 100, 5097—5102 (2003). 49) Yokoyama K., Ihara M., Ebihara S., Suzuki M., Proc. Jpn. Acad., 81B, 463—470 (2005). 50) Peeters E., Thia-Toong T.-L., Gigot D., Maes D., Charlier D., Mol. Microbiol., 54, 321—336 (2004). 51) Brinkmann A. B., Bell S. D., Lebbink R. J., de Vos W. M., van der Oost J., J. Biol. Chem., 277, 29537—29549 (2002). 52) Koike H., Amano N., Tateno M., Ohfuku Y., Suckow J. M., Suzuki M., Proc. Jpn. Acad., 75B, 37—42 (1999). 53) Suzuki M., Amano N., Koike H., Proc. Jpn. Acad., 79B, 92—98 (2003). 54) Dartois V., Liu J., Hoch J. A., Mol. Microbiol., 25, 39—51 (1997). 55) Beloin C., Exley R., Mahé A.-L., Zouine M., Cubasch S., Le Hégarat F., J. Bacteriol., 182, 4414—4424 (2000).

186 56) Belitsky B. R., Gustafsson M. C. U., Sonenshein A. L., Von Wachenfeldt C., J. Bacteriol., 179, 5448—5457 (1997). 57) Suzuki M., Aramaki H., Koike H., Proc. Jpn. Acad., 79B, 242—247 (2003). 58) Stainer R. Y., Palleroni N. J., Doudoroff M., J. Gen. Microbiol., 43, 159—271 (1996). 59) Segerer A., Langworthy T. A., Stetter K. O., Syst. Appl. Microbiol., 10, 161—171 (1988). 60) Kawashima T., Yokoyama K., Higuchi S., Suzuki M., Proc. Jpn. Acad., 81B, 204—219 (2005). 61) Keuntje B., Masepohl B., Klipp W., J. Bacteriol., 177, 6432—6439 (1995). 62) Cho K., Winans S. C., Mol. Microbiol., 22, 1025—1033 (1996).

Vol. 31, No. 2 63) Lodwig E., Kumar S., Allaway D., Bourdes A., Prell J., Priefer U., Poole P., J. Bacteriol., 186, 842—849 (2004). 64) Napoli A., van der Oost J., Sensen C. W., Charlebois R. L., Rossi M., Ciaramella M., J. Bacteriol., 181, 1474—1480. 65) Yokoyama K., Suzuki M., Proc. Jpn. Acad., 81B, 129—139 (2005). 66) Kyrpides N. C., Ouzounis C. A., Trends Biochem. Sci., 20, 140—141 (1995). 67) Suzuki M., J. Mol. Biol., 207, 61—84 (1989). 68) Suzuki M., EMBO J., 8, 797—804 (1989). 69) Suzuki M., Proc. Jpn. Acad., 81B, 403—409 (2005). 70) Yokoyama K., Ihara M., Ebihara S., Suzuki M., Proc. Jpn. Acad., 82B, 33—44 (2006). 71) Wang Q., Calvo J. M., J. Mol. Biol., 229, 306—318 (1993).