Vol. 171, No. 6JOURNAL OF BACTERIOLOGY, June 1989, p. 3494-35030021-9193/89/063494-10$02.00/0Copyright © 1989, American Society for Microbiology

Nucleotide Sequence of the FNR-Regulated Fumarase Gene (fumB)of Escherichia coli K-12

PHILIP J. BELL,t SIMON C. ANDREWS, MIRTA N. SIVAK,t AND JOHN R. GUEST*

Department of Microbiology, University of Sheffield, Sheffield S10 2TN, United Kingdom

Received 25 January 1989/Accepted 16 March 1989

The nucleotide sequence of a 3,162-base-pair (bp) segment ofDNA containing the FNR-regulatedfumB gene,which encodes the anaerobic class I fumarase (FUMB) of Escherichia coli, was determined. The structural genewas found to comprise 1,641 bp, 547 codons (excluding the initiation and termination codons), and the geneproduct had a predicted M, of 59,956. The amino acid sequence of FUMB contained the same number ofresidues as did that of the aerobic class I fumarase (FUMA), and there were identical amino acids at all but 56positions (89.8% identity). There was no significant similarity between the class I fumarases and the class II

enzyme (FUMC) except in one region containing the following consensus: Gly-Ser-Xxx-Ile-Met-Xxx-Xxx-Lys-Xxx-Asn. Some of the 56 amino acid substitutions must be responsible for the functional preferences of theenzymes for malate dehydration (FUMB) and fumarate hydration (FUMA). Significant similarities between thecysteine-containing sequence of the class I fumarases (FUMA and FUMB) and the mammalian aconitases weredetected, and this finding further supports the view that these enzymes are all members of a family ofiron-containing hydrolyases. The nucleotide sequence of a 1,142-bp distal sequence of an unidentified gene(genF) located upstream offumB was also defined and found to encode a product that is homologous to theproduct of another unidentified gene (genA), located downstream of the neighboring aspartase gene (aspA).

Fumarase or fumarate hydratase (FUM; S-malate hydro-lyase; EC 4.2.1.2) catalyzes the interconversion of fumarateand L-malate. Gene-cloning and enzymological studies haverecently shown that Escherichia coli K-12 possesses at leastthree fumarase genes (fumA, fumB, and fumC), each en-coding a different fumarase: FUMA, FUMB, and FUMC (7,8, 40). ThefumA andfumC genes are adjacent at 35.5 min inthe linkage map, whereas thefumB gene is located close tothe mel operon at 93.5 min. The three enzymes representtwo biochemically distinct classes of fumarase, class I andclass II, which probably correspond to two general classes ofhydrolyase (40). The class I enzymes, FUMA and FUMB,are typically thermolabile homodimeric enzymes of Mr120,000 (2 x 60,000), and it has been suggested that they aremembers of a family of iron-dependent hydrolyases whichincludes aconitase and maleate hydratase (17, 40). Indeed, ithas recently been shown that the activity of the labile FUMAof E. coli W can be restored by anaerobic incubation withferrous ions and thiol compounds (43). The fumA gene hasbeen sequenced (16), and it is sufficiently homologous to thefumB gene for hybridization to occur under high-stringencyconditions (7). The class II enzymes are typified by FUMC,which is a thermostable enzyme of native Mr 200,000 com-prising four identical subunits (4 x 50,000). The fumarases ofBacillus subtilis, Saccharomyces cerevisiae, and mammaliansources are all class II enzymes having a remarkably highdegree of sequence identity with FUMC (.60%) and withthe aspartases (38%) and argininosuccinases (15%) fromseveral sources (17, 20, 38-40). However, there is no signif-icant homology between the class I (FUMA) and class IIfumarases except for one segment containing a Gly-Ser-Xxx-Ile-Met-Xxx-Xxx-Lys-Xxx-Asn consensus.

* Corresponding author.t Present address: ICI Seeds, Jealotts Hill Research Station,

Bracknell, Berkshire RG12 6EY, United Kingdom.t Present address: Robert Hill Institute, University of Sheffield,

Western Bank, Sheffield S10 2TN, United Kingdom.

The physiological functions of the three E. coli fumaraseshave been investigated with mutants defective in one ormore of the genes, with transformants containing multicopyplasmids expressing each of the genes, and with strainscontaining single-copy fum-lacZ translational fusions inwhich a hybrid P-galactosidase is expressed from each of thefum promoters (10, 37, 40). These investigations have shownthat multiple copies of each of the genes complement thenutritional lesions (affecting aerobic respiratory growth onfumarate or anaerobic respiratory growth on glycerol plusmalate) that are exhibited by different fum mutants. How-ever, it is clear that FUMA is a citric acid cycle enzyme,because it is strongly expressed under aerobic conditions butrepressed by glucose and anaerobiosis. The anaerobic re-pression is presumably mediated by the arcA (dye)-encodedregulator of aerobic metabolism (11). This contrasts withFUMB, which is strongly expressed during anaerobic respi-ratory growth with glycerol plus fumarate but is otherwiseonly weakly induced by anaerobiosis and relatively insensi-tive to catabolite repression. Furthermore, the anaerobicexpression ofFUMB is reduced infnr mutants and amplifiedby a multicopy fnr plasmid (37), which is consistent withFUMB being an anaerobic enzyme requiring the anaerobictranscription factor (FNR; 23, 24) for expression. The rela-tive affinities of the two enzymes for fumarate and malate arealso consistent with specific aerobic and anaerobic func-tions, because FUMA has the properties of a fumaratehydratase, whereas FUMB is a malate dehydratase (40).Thus, it appears that E. coli contains two analogous butdifferentially regulated class I fumarases, specificallyadapted for reciprocal roles in the overall oxidation offumarate in the citric acid cycle (FUMA) or in the provisionof fumarate as an anaerobic electron acceptor (FUMB). Thissituation closely parallels that of the specific and differen-tially regulated succinate dehydrogenase (aerobic) and fuma-rate reductase (anaerobic) complexes in E. coli. The precisefunction of FUMC is not known, although its greater affinityfor fumarate than malate resembles that of FUMA, the

3494

on August 20, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

FUMARASE B GENE OF E. COLI 3495

genF fumB

HB9A Pi P Bc T E

melB melA

_ A H X P Bc H

PIV

P Evl

; PI1 $3 64B

5 6 7 a

PIB,,I

F

genFH , Bg, .........

Ev,H Bg

pGS56 ,pv

i pGS93fum - - -

P Bc Sp HpIF

0 ~--

0.-~

FIG. 1. Restriction map of thefumB-mel region at 93.5 min in the W. coli chromosome and summary of DNA sequence data obtained fromM13 clones. Polarities of the genes are indicated; left to right corresponds to clockwise in the linkage map, and the scales are in kilobases.The plasmid subclones are Apr Tcs derivatives of pBR322 (7 and 8); open bars represent bacterial DNA. Symbols: -, -, positions and extentsof sequencing for M13 clones; 0, sequences obtained with specific primers; , fumB coding region and part of the unidentified genFcoding region. Relevant restriction sites are abbreviated as follows: Aval, A; BamHI, B; BclI, Bc; BglII, Bg; EcoRI, E; HindIII, H; HpaI,Hp; PstI, P; PvuI, PI; PvuII, Pll; SphI, Sp. Vector sites are indicated by a subscript v.

aerobic fumarate hydratase. FUMC represents the stablefumarase activity observed under all growth conditions,accounting for up to 25% of the fumarase activity in aerobicorganisms. ThefumC promoter is relatively weak and insen-sitive to environmental stimuli, and it is conceivable thatfumC expression is linked in some way to that of theupstreamfumA gene (37).

This paper reports the nucleotide sequence of the fumBgene and the amino acid sequence of its product, FUMB. Ithighlights the very high degree of sequence homology be-tween the class I fumarases (FUMA and FUMB) at both theDNA and protein levels. Significant homologies were alsodetected between the cysteine-containing sequences of ac-onitase and the class I fumarases. A partial sequence wasobtained for an unidentified gene (genF) that is adjacent tofumB, and its product was found to be very similar to that ofanother unidentified gene (genA) lying next to the aspartasegene (aspA) just 17 kilobase pairs (kbp) from genF in the E.coli chromosome.

MATERIALS AND METHODS

Bacterial strains, plasmids, and bacteriophages. E. coliGM242 (dam-3 recAl) was used as a transformable host forplasmid construction and preparing Bc/I-susceptible plasmidDNA (14), and JM101 [thi supE A(proAB-lac) (F' traD36proAB+ lacZAM15 laclq)] was used for propagating M13derivatives (15). Plasmid pGS93 (fumB+) was the source ofDNA for sequencing the fumB gene (Fig. 1). It was derivedby recloning a 3.85-kbp HindIII-BamHI fragment (H-B)from pGS56 (fumB+-malAB+) into the corresponding sitesof pBR322 (7). The replicative forms of M13mpl8 andM13mpl9 were used as vectors for preparing DNA tem-plates for sequence analysis (41).

Nucleotide sequence analysis. The sequencing strategy in-volved cloning specific fragments of pGS93 obtained withdifferent combinations of EcoRI, BamHI, BclI, BglII, PstI,

and SphI into the corresponding sites of M13mpl8 andM13mp19 (Fig. 1). Single-stranded M13 DNA templateswere sequenced by the dideoxy-chain termination method,using two universal primers, [a-35S]thio-dATP, and buffergradient gels with DNA polymerase (Klenow fragment) orSequenase (2, 21, 32). The amount of sequence obtainedfrom some clones was increased by using specific oligonu-cleotide primers designed on the basis of the initial se-quences or on the fumA sequence (once the very closehomology between fumB and fumA had become apparent).The oligonucleotide primers were made with an AppliedBiosystems 381A DNA synthesizer. Nucleotide sequenceswere compiled and analyzed with the Staden computerprograms (26, 27, 29, 30), and sequence comparisons weremade by using the DIAGON (28) and DAPSEARCH (3, 4)programs.

Materials. Restriction endonucleases and T4 DNA ligasewere purchased from Bethesda Research Laboratories, Inc.(Gaithersburg, Md.). DNA polymerase (Klenow fragment)and Sequenase were purchased from Boehringer MannheimBiochemicals (Indianapolis, Ind.) and Cambridge Bio-Science, respectively. The M13mpl8 and 19 replicative-formDNAs were from Pharmacia, Inc. (Piscataway, N.J.).

RESULTS AND DISCUSSION

Nucleotide sequence of the fumB region. The fumB geneencoding the FUMB monomer (approximate Mr, 61,000)was previously located in the 2.5-kbp BglII-SphI segment ofpGS56 and pGS93 and shown to have the polarity illustratedin Fig. 1 (7, 8). The complete nucleotide sequence of anoverlapping 3,162-bp HindIII-HpaI fragment was deter-mined (Fig. 1 and 2). The sequencing strategy involvedcloning specific restriction fragments in M13 and sequencingthe inserts with the aid of several synthetic oligonucleotideprimers (Fig. 1). The nucleotide sequence (Fig. 2) was fully

. s i i i i.1 .-a i I

- i ====:j I - -

1,

VOL. 171, 1989

J'

I

on August 20, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

geniF ( 20)

A%";C-TCZGGCGGTCTTGATGTCATGCTGCAAATTGCCGAGAAGCTGCTGCGCCGCAACCCGAAATATGTCTCAATTGTCGCGCCGTTTG10 20 30 40 50 60 70 80 90

(40)V T7 C T L T I L C G T G H V V Y T I L P I I Y D V A I K N N

TGACCTGTACACTGACCATTCTTTGCGGTACGGGTCATGTGGTTTACACCATTCTGCCGATCATCTACGACGTCGCCATTAAGAACAACA100) 110 120 130 140 150 160 170 180

(,C) (80)

I R P E R P M A A S S I G A QM G I I A S P V S V A V V S L

TCCGTCCGGAACGTCCGATGGCGGCAAGTTCTATCGGTGCACAGATGGGGATTATCGCCAGTCCGGTGTCGGTTGCGGTCGTGTCTCTGG190 200 210 220 230 240 250 260 270

(100)V A M L G N V T F D G R H( L E F L D L L A I T I P S T L I G

TTGCGATGCTGGGTAATGTCACCTTTGATGGTCGCCATCTTGAGTTCCTCGA'I;TGCTGGCAATCACCATTCCATCGACGTTAATCGGTA280 290 300 310 320 330 340 350 360

(120) rPvuI rPVUII1 -BglII1 (140)I L I I FS W1 F R G lDL D K D E E F Q K F I S V P E NTCCTGGCGATCGGTATCTTCAGCTGGTTCCGCGGTAAAGATCTGGATAAAGACGAAGAGTTCCAGAAATTCATCTCCGTACCGGAAAACC

370 380 390 400 410 420 430 440 450(160) ArvaI

R E Y V Y G D T A T L L D K K L P K S N W L A M W I F L G AGTGAGTATGTTTACGGTGATACCGCGACGCTGCTGGATAAAAAACTGCCGAAAAGCAACTGGCTGGCAATGTGGATTTTCCTCGGGGCAA

460 470 480 490 500 510 520 530 540(180) (200)

I A V V A L L G A D S D L R P S F G G K P L S M V L V I Q MTCGCTGTAGTCGCCCTTCTTGGTGCTGATTCGGACCTGCGTCCATCCTTCGGCGGCAAACCGCTGTCGATGGTACTGGTTATTCAGATGT

550 560 570 580 590 600 610 620 630(220)

F M L L T G A L I I I L T K T N P A S I S K N E V F R S G MTTATGCTGCTGACCGGGGCGCTGATTATTATCCTGACCAAAACCAATCCCGCGTCTATCTCAAAAAACGAAGTCTTCCGTTCCGGTATGA

640 650 660 670 680 690 700 710 720(240) (260)

I A I V A V Y G I A W M A E T M F G A H M S E I Q G V L G ETCGCCATCGTGGCGGTGTACGGTATCGCATGGATGGCAGAAACCATGTTCGGTGCGCATATGTCTGAAATTCAGGGCGTACTGGGTGAAA

730 740 750 760 770 780 790 800 810(280)

M V K E Y P W A Y A I V L L L V S K F V N S Q A A A L A A ITGGTGAAAGAGTATCCGTGGGCCTATGCCATTGTTCTGCTGCTGGTTTCCAAGTTTGTAAACTCTCAGGCTGCGGCGCTGGCGGCGATTG

820 830 840 850 860 870 880 890 900

(300) rPVUI- (320)V P V A L A I IG V D P A Y I V A S A P A C Y G Y Y I L P T Y

TTCCGGTCGCGCTGGCGATCGGCGTTGATCCGGCATACATCGTGGCTTCAGCACCGGCTTGCTACGGTTATTACATCCTGCCGACTTATC910 920 930 940 950 960 970 980 990

(340)P S D L A A I Q F D R S G T T H I G R F V I N H S F I L P GCCAGCG.' T CTGGCAGCGATTCAGTT TGACCCTTCCGGCAC rC.CCCACATCGGTCGCTTCGTCATCAACCACAGCTTTATTCTGCCGGGGT

1000 1010 1020 1030 1040 1050 1060 1070 1080360) (380) -3L I G V S V S C V F G W I F A A M Y G F L * 3

TGATTGGTGTGAGCGTATCGTGCGTCTTCGGCTGGATCTTCGCCGCGATGTACGGGTTCTTATAA ATGCACTTTGCGTGCCGCCCGTGAC1090 1100 1110 1120 1130 1150 1i60 1170

-10 1 fumB 10> _ fM S N K P F IY Q A P F PTACGCGGCACGCCATTTTCGAATAACAAATACAGAGTTACEgSETGGAAGCTATGTCAAACAAACCCTTTATCTACCAGGCACCTTTCCC

1180 1190 1200 1210 1220 1230 1240 1250 - 126020 30 40

M G K D N T E Y Y L L T S D Y V S V A D F D G E T I L K V EGATGGGGAAAGACAATACCGAATACTATCTACTCACTTCCGATTACGTTAGCGTTGCCGACTTCGACGGCGAAACCATCCTGAAAGTGGA

1270 1280 1290 1300 1310 1320 1330 1340 135050 60 70

P E A L T L V A Q Q A F H D A S F M L R P A H Q K Q V A A IACCAGAAGCCCTGACCCTCGTGGCGCAGCAAGCCTTTCAC(iACGCTTCTTTTATGCTCCGCCCGGCACACCAGAAACAGGTTGCGGCTAT

1360 1370 1380 1390 1400 1410 1420 1430 144080 90 100

L H D P E A S E N D K Y V A L Q F L R N S E I A A K G V L PTCTTCACGATCCAGAAGCCAGCGAAAACGACAAGTACGTGGCGCTGCAATTCTTAAGAAACTCCGAAATCGCCGCCAAAGGCGTGCTGCC

1450 1460 1470 1480 1490 1500 1510 1520 1530110 120 130

T C Q D T G T A I I V G K K G Q R V W T G G G D E E 7 L S KGACCTGCCAGGATACCGGCACCGCGATCATCGTCGGTAAAAAAGGCCAGCGCGTGTGGACCGGCGGCGGTGA.TGAAGAAACGCTGTCGAA

1540 1550 1560 1570 1580 1590 1600 1610 1620140 150 160

G V Y N T Y I E D N L R Y S Q N A A L D M Y K E V N T G T NAGGCGTCTATAACACCTATATCGAAGATAACCTGCGCTATTCACAGAATGCGGCGCTGGACATGTACAAAGAGGTCAACACCGGCACTAA

1630 1640 1650 1660 1670 1680 1690 1700 1710170 180 190

L P A Q I D L Y A V D G D E Y K F L C V A K G G G S A N K TCCTGCCTGCGCAAATCGACCTGTACGCGGTAGATGGCGATGAGTACAAATTCCTTTGCGTTGCGAAAGGCGGCGGCTCTGCCAACAAAAC

17220 1730 1740 1750 1760 1770 1780 1790 1800FIG. 2. Nucleotide sequence of the fumB gene and primary structure of its product. The nucleotide coordinates are assigned relative to

the first base of the HindIII site. The primary structure predicted for FUMB is shown above the nucleotide sequence. The amino acidsequence of the C-terminal segment of an unidentified reading frame, genF, is shown, with residue numbers in brackets to indicate that theN-terminal residues are not included. A putative promoter is indicated at the -35 and -10 sequences; El, putative transcription start site.Three potential FNR-binding sites and two half-sites are underlined (see text). Key restriction sites, stop codons (*), a potentialribosome-binding site (boxed), and significant regions of hyphenated dyad symmetry (converging arrows) are also indicated.

3496

on August 20, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 171, 1989 FUMARASE B GENE OF E. COLI 3497

200 210C jPst I

Y L Y 4 E T K A L L T P G K L K N F L V E K FR T L G ITAFTGTATCTCTACCAGGAAACCAAAGCCCTGCTGACTCCCGGCAAACTGAAAAACTTCCTCGTCGAGAAAATGCGTACCCTCGGTACTGCAGC

181C 1820 1830 1840 1850 1860 1870 1880 1890

230 240 250C P P Y H I A F V I G G T S A E T N L K T V K L A S A H Y Y

CTGCCCGCCGTACCATATCGCGTTTGTGATTGGCGGTACGTCTGCGGAAACCAACCTGAAAACCGTCAAGTTAGCAAGCGCTCACTATTA1900 1910 1920 1930 1940 1950 1960 1970 1980

260 rPVUIIi 280D E L P T E G N E H G Q A F R D V| Q LIE Q E L L E E A Q K L

CGATGAACTGCCGACGGAAGGGAACGAACATGGTCAGGCGTTCCGCGATGTCCAGCTGGAACAGGAACTGCTGGAAGAGGCCCAGAAACT1990 2000 2010 2020 2030 2040 2050 2060 2070

290 300 310G L G A Q F G G K Y F A H D I R V I R L P R H G A S C P V G

CGGTCTTGGCGCGCAGTTTGGCGGTAAATACTTCGCGCACGACATTCGCGTTATCCGTCTGCCACGTCACGGCGCATCCTGCCCGGTCGG2080 2090 2100 2110 2120 2130 2140 2150 2160

320 330 340M G V S C S A D R N I K A K I N R E G I W I E K L E H N P G

CATGGGCGTCTCCTGCTCCGCTGACCGTAACATTAAAGCGAAAATCAACCGCGAAGGTATCTGGATCGAAAAACTGGAACACAACCCAGG2170 2180 2190 2200 2210 2220 2230 2240 2250

350 360 370Q Y I P Q E L R Q A G E G E A V K V D L N R P M K E I L A Q

CCAGTACATTCCACAAGAACTGCGCCAGGCCGGTGAAGGCGAAGCGGTGAAAGTTGACCTTAACCGCCCGATGAAAGAGATCCTCGCCCA2260 2270 2280 2290 2300 2310 2320 2330 2340

380 390 400L S Q Y P V S T R L S L T G T E I V G R D I A H A K L K E L

GCTTTCGCAATACCCGGTATCCACTCGTTTGTCGCTCACCGGCACCATTATCGTGGGCCGAGATATTGCACACGCCAAGCTGAAAGAGCT2350 2360 2370 2380 2390 2400 2410 2420 2430

410 420 430

I D A G K E L P Q Y I K D H P I Y Y A G P A K T P A G Y P SGATTGACGCCGGTAAAGAACTTCCGCAGTACATCAAAGATCACCCGATCTACTACGCGGGTCCGGCGAAAACCCCTGCCGGTTATCCATC

2440 2450 2460 2470 2480 2490 2500 2510 2520

440 450 FBclI} 460G S L G P T T A G R M D S Y V D L L Q S H G G S M I L A K

AGGTTCACTTGGCCCAACCACCGCAGGCCGTATGGACTCCTACGTGGATCTGCTGCAATCCCACGGCGGCAGCATGATCATGCTGGCGAA2530 2540 2550 2560 2570 2580 2590 2600 2610

470 480 490G N R S Q Q V T D A C H K H G G F Y L G S I G G P A A V L A

AGGTAACCGCAGTCAGCAGGTTACCGACGCGTGTCATAAACACGGCGGCTTCTACCTCGGTAGCATCGGCGGTCCGGCGGCGGTACTGGC2620 2630 2640 2650 2660 2670 2680 2690 2700

500 510 520Q Q S I K H L E C V A Y P E L G M E A I W K I E V E D F P A

GCAGCAGAGCATCAAGCATCTGGAGTGCGTCGCTTATCCGGAGCTGGGTATGGAAGCTATCTGGAAAATCGAAGTAGAAGATTTCCCGGC2710 2720 2730 2740 2750 2760 2770 2780 2790

530 540 547F I L V D D K G N D F F Q Q I V N K Q C A N C T K *

GTTTATCCTGGTCGATGACAAAGGTAACGACTTCTTCCAGCAAATCGTCAACAAACAGTGCGCGAACTOCACTAAGTAACC'.CTTCGGC C

2800 2810 2820 2830 2840 2850 2860 2870 2880

rS phI-ICAGCGCCTGGCAGCATGCTGCCAGGTGATCCCC CTGGCCACCTCTTTTGCGATTGTAATTTCACGCTTGCTGGTGAATAGTCAGTATTTT

2 890 ' R2900 ' 2910 2 9290 2930 2940 2950 2960 2970

CCCTGATTTGCGAACTCACCATGACCCGCACACTCAAGCCGTTAATTCTTAACACCAGCGCACTGACGCTAACGTTAATCCTGATTTATA2980 2990 3000 3010 3020 3030 3040 3050 3060

CCGGCATTTGGCCCATGACAAACTCACCTGGCTGATGGAAGTGACACCGGTGATTATTGTCGTGCAGCTACTGCTTGCCACCGCCAGACG3070 3080 3090 3100 3110 3120 3130 3140 3150rHpa I-)

TTATCCGTTAA,3160

overlapped, and 90% (100% offumB) was from both DNAstrands. Two potential coding regions were detected with theFRAMESCAN program (30). One was identified as thefumB gene because it encodes a polypeptide of Mr 59,956(547 amino acid residues) that is highly homologous with thefumA gene product (Fig. 2, positions 1226 to 2866). Theother, designated genF, had the same polarity as fumB andencodes some 380 amino acid residues at the C-terminal endof an unidentified gene product, GP-F (Fig. 2, positions 1 to1142). There was no evidence of a coding region correspond-ing to the C-terminal end of the product of the putative distalgene of the mel operon, melC (42), in the 290-bp sequencedsegment of the melB-fumB intergenic region (approximately790 bp).

Features of the nucleotide sequence. The fumB coding

region is preceded by a potential ribosome-binding site (Fig.2), and the proposed translational start site gave a relativelyhigh score according to the PERCEPTRON algorithm ofStormo et al. (31). ThefumB coding region is 79% identicalto that offumA. Of the 340 differences in the coding region,239 are silent mutations that affect synonymous codons for231 residues, and the majority (220) represent single changesat the third position. The remaining 101 differences affect 56codons at one (21), two (25), or three (10) positions. Thecodon usage of the fumB gene is shown in (Fig. 3). Thefrequency of modulatory codons (0.9%) is extremely low,and the distribution of optimal energy codons (56%) in thediagnostic set suggests that fumB is moderately expressed(6). The fumB gene may therefore have a somewhat higherpotential for expression than do the other fum genes, for

on August 20, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

3498 BELL ET AL.

fumA B

UUU F 6 6UUC F 10 11UUA L 3 2UUG L 5 1

CUU L 8 7CUC L 5 10*CUAL 1 1CUG L 29 29

AUU I 12 8AUC I 18 23*AUA I 1 0AUG M 11 11

GUU VGUC VGUA VGUG V

9 76 113 4

11 9

fumA B fumA B

UCU S 6 3 UA[ Y 16 8UCC S 5 8 UAC Y 10 18UCA S 1 4 UAA - 1 1UCG S 5 3 UAG - 0 0

ICCU P 2 3CCP 3 2

CCA P 7 7CCG P 15 15

ACU T 4 6ACC T 10 19ACA T 2 0ACG T 11 4

UGU C 6 1UGC C 3 8UGA - 0 0

UGG W 3 3

CAU H 5 4 CGU RCAC H 7 11 CGC RCAA Q 9 7 *CGA RCAG

Q20 23 *CGG R

AAU N 10 21 AGU SAAC N 11 19 AGC SAAA K 34 33 *AGA RAAG K 4 5 *AGG R

GCU A 6 6 GAU D 21 14GCC A 12 1S GAC D 8 15GCA A 5 7 GAA E 30 28GCG A 29 25 GAG E 8 8

9 6

12 90 12 0

GGU G 23 18GGC G 15 30

*GGA G 7 0*GGG G 6 2

FIG. 3. Codon usage in the fumA and fumB genes of E. coli.Boxed codon pairs are those whose use varies between strongly andweakly expressed genes; *, potential modulatory codons (6). Initi-ation codons are included with the methionine codons for thefumA(16) and fumB sequences.

which the corresponding values are 3.0 and 51% (fumA) and3.9 and 46% (fumC) (16, 39).A putative promoter sequence was detected in the rela-

tively short 77-bp genF-fumB intergenic region by using theANALYSEQ program (29) (Fig. 2). The putative -35 regionis flanked by a region of hyphenated dyad symmetry (AG,24.8 kcal [ca. -104.8 KJ] mol-1; 34), which could functionas the transcriptional terminator for genF because it islocated 10 bp downstream of the genF stop codon and isclosely followed by a run of four T's (19). The expression offumB is activated under anaerobic conditions by the anaer-

obic transcriptional regulator, FNR (37), and potential FNR-binding sites were sought by using a 12-bp symmetricalconsensus, TGATA- -TATCAA, based on that proposedby Spiro and Guest (25; S. Spiro, M. H. Todd, P. J.Artynfliuk, and J. R. Guest, Biochem. Soc. Trans. 16:755-756, 1988). The best match was located in the genFcoding region; it had identities at eight positions, TTGAIT- -

TGTGAG (1080 to 1093 in Fig. 2). Two further sequenceswith slightly less homology were located closer to thefumBpromoter at the end of the genF coding region, CGGGTT- -

TATAAA (1133 to 1146) and TTCTIA- -AAIGCA (1137 to1150), and there were also two FNR half-sites, CATCAA(1052 to 1057) and TATCGT (1096 to 1101). The significanceof the potential FNR sites is difficult to assess, partlybecause their homologies with the consensus were notparticularly strong. This poor quality may be reflected in therelatively weak anaerobic induction offumB. Nevertheless,one or more of the sites may be functional, and their locationin genF may be unimportant if genF is repressed under theconditions that induce fumB expression. Alternatively, it isconceivable that genF andfumB are both expressed from anFNR-regulated promoter situated upstream of genF or thatFNR exerts an indirect effect on fumB expression. Apartfrom a cyclic AMP receptor protein (CRP) half-site, TGTGA(1088 to 1093), no sequences resembling the CRP-bindingsite consensus were detected in a comparable search of thefumB promoter region. This is consistent with the observa-tion that fumB is not subject to catabolite repression (37).The sequence contained two significant regions of dyad

symmetry immediately downstream of the fumB structuralgene (Fig. 2). The first was a 20-bp palindrome centered onthe SphI testriction site (AG, -18.8 kcal [ca. -78.7 KJ]mol- 1; 34). The second should be capable of forming a stable

stem-loop structure in the mRNA transcript (AG, -15.4 kcal[ca. -64.4 KJ] mol-'; 34), although it overlapped thepalindromic sequence. There was also a T-rich sequencesome 7 bp further downstream, which suggests that one orboth of these potential secondary structures function as arho-independent terminator forfumB (19).Comparison between FUMB and FUMA. The fumB gene

was found to encode a product of Mr 59,956. This value isreasonably close to that of 61,000 predicted for FUMB fromits electrophoretic mobility in denaturing polyacrylamide gel(7). The FUMB monomer is only slightly smaller than theFUMA monomer (Mr, 60,163). The two proteins contain thesame number of amino acid residues (547, excluding theinitiating formylmethionine) and have very similar aminoacid compositions (Fig. 3). The major differences involvecharged residues, FUMB having six fewer arginine residuesand two fewer glutamate residues. However, the net differ-ence of four negative charges in FUMB is to some extentoffset by the presence of three additional histidine residuesin FUMB.The FUMA and FUMB amino acid sequences exhibited a

remarkable 89.8% sequence identity, and the similarityincreased to 94.0% when conservative substitutions scoring.0.1 in the mutation data matrix, MDM78 (3), were included.The amino acid replacements in FUMB relative to theFUMA sequence are highlighted in Fig. 4. The replacementsare clustered in several regions, notably the N-terminalsegment (residues 1 to 51), the C-terminal segment (residues538 to 547), and some other segments such as residues 270 to281, 330 to 359, and 402 to 408. These account for 19 of the25 nonconservative substitutions, and it is also apparent thattwo of these segments are significantly more acidic in FUMB(residues 1 to 51 and 330 to 359 each lose two net positivecharges). The hydropathy profiles of both proteins are typi-cal of soluble proteins, and the small differences betweenFUMA and FUMB are entirely consistent with the aminoacid substitutions.A putative active-site decapeptide of FUMA is totally

conserved in the FUMB sequence: Gly-Ser-Met-Ile-Met-Leu-Ala-Lya-Gly-Asn, residues 455 to 464 (Fig. 4). Thissequence was detected in FUMA as the only region showingsignificant homology with FUMC and with other members ofthe class II fumarase-aspartase-argininosuccinase family ofenzymes (38, 40). Indeed, the five underlined residues areinvariant in eight independent sequences representing thethree types of enzyme. They include a putative active-sitemethionine residue (Met-459 in FUMA and FUMB) and alysine residue (Lys-462 in FUMA and FUMI3) that may beimportant in binding the dicarboxylic acid substrates. Thehigh degree of sequence conservation in the decapeptidestrongly suggests that the corresponding residues performsimilar functions in both the class I and class II fumarases.Each of the 9 cysteine residues is conserved between FUMAand FUMB; 9 of the 10 methionine and 10 of the 12 or 15histidine residues are likewise conserved (Fig. 4).The very high degree of sequence similarity between

FUMA and FUMB is indicative of a very close evolutionaryrelationship. They would appear to have emerged by geneduplication and divergent evolution of a common ancestralgene. Their functional diversity is largely imposed by thedifferentially regulated promoters such that FUMA is aCRP-dependent aerobic enzyme, whereas FUMB is anFNR-dependent anaerobic enzyme. However, the enzymesare themselves specifically adapted for their functionallycompartmented metabolic roles. The net effect of the 56amino acid differences between the two enzymes is to

J. BACTERIOL.

on August 20, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

FUMARASE B GENE OF E. COLI 3499

FUMA

FUMB

1 SNKPFHYQAPFPLKKDDTEYYLLTSEHVSVSEFEGQEILKVAPEALTLLARQAFHDASFM 60:..... MG....+:+ +: ++ + +: :. + +... I . MG. .N . DYA.D.E......... .E . V.Q

FUMA 61 LRPAHQQQVADILRDPEASENDKYVALQFLRNSDIAAKGVLPTCQDTGTAIIVGKKGORV 120

FUMB K.. .A..H....E

FUMA 121 WTGGGDEAALARGVYNTYIEDNLRYSONAPLDMYKEVNTGTNLPAQIDLYAVDGDEYKFL 180:F++. +

FUMB.... . .ET.SK.A...................A

FUMA

FUMB

FUMA

FUMB

181 CIAKGGGSANKTYLYGETKALLTPGKLKNYLVEKMRTLGTAACPPYHIAFVIGGTSAETN 240+F +

.V . . . . . . . . . . . . . F . . . . . . . . . . . . . .

241 LKTVKLASAKYYDELPTEGNEHGQAFRDVELEKELLIEAQNLGLGAQFGGKYFAHDIRVI 300:........H................... Q .E ...K .................+.+

H.Q..Q...E...K.....

FUMA 301 RLPRHGASCPVGMGVSCSADRNIKAKINROGIWIEKLEHNPGKYIPEELRKAGEGEAVRV 360+ : + :+

FUMB.E........ .E.Q K.

FUMA 361 DLNRPMKEILAOLSQYPVSTRLSLNGTIIVGRDIAHAKLKERMDNGEGLPQYIKDHPIYY 420

FUMB ........................ ................LI.A.KE.

FUMA 421 AGPAKTPEGYASGSLGPTTAGRMDSYVDQLQAQ X jSMIMLAKGNSOOVTDACKKHGGFY 480+ : ++:

FUMB .A.P.L. .SH .H.

FUMA 481 LGSIGGPAAVLAQGSIKSLECVEYPELGMEAIWKIEVEDFPAFILVDDKGNDFFQOIQLT 540

FUMB . ... .. ...A. VNK

FUMA 541 OCTRCVK 547

FUMB ..AN.T.FIG. 4. Amino acid sequence of FUMA, highlighting the differences found in the FUMB sequence. Symbols: +, conservative

substitutions scoring 0.1 in the MDM7. matrix (28); :, nonconservative substitutions. The putative active-site decapeptide detected in theclass I and class II fumarases and in related enzymes is boxed, and the invariant residues are underlined.

decrease the affinity of FUMA for fumarate (11-fold) andincrease its affinity for malate (1.8-fold), thus transformingthe fumarate hydratase (FUMA) into a malate dehydratase(FUMB) or vice versa. It would be interesting to constructhybrid enzymes or enzymes with one or more substitutionsin order to define the structural basis for these functionaladaptations and possibly to create enzymes with intermedi-ate properties.

It has previously been suggested that the class I fumarasesbelong to a family of iron- and sulfur-containing carboxylicacid hydrolyases which includes aconitase, maleate dehy-dratase, and hydroxyglutaryl-coenzyme A dehydratase (17,22, 40). Evidence supporting this view comes from therelatively high instability of these enzymes, their high cys-teine contents (1.5% by weight), and the presence of cystei-nyl-prolyl sequences. The latter are relatively infrequentexcept in ferredoxins and other proteins containing Fe-Sclusters, in which the cysteine residues serve as ligands forthe Fe atoms of the clusters. Further support has now beenobtained by demonstrating that inactive FUMA can bereactivated by Fe2+ and thiol reagents under anaerobic

conditions and by showing that purified FUMA is brown andcontains a characteristic shoulder at 280 to 420 nm in itsabsorption spectrum (43; D. H. Flint, personal communica-tion).The sequences of several tryptic peptides containing

seven of the estimated 11 or 12 cysteine residues in beefheart aconitase have recently been added to that of a singlepeptide containing the reactive cysteine residue of pig heartaconitase (9, 18). A linear compilation of the aconitasesequences was compared with the complete FUMA andFUMB sequences by using the DIAGON program (28).Several significant homologies (double-matching probability,<0.001; span, 9 to 13) were detected, primarily with thecysteine-containing regions of the fumarases (Fig. 5). Thesimilarities with the pig heart peptide are primarily due to thepresence of the cysteinyl-prolyl dipeptide, which it shareswith the two cysteiny-prolyl-containing regions in the fuma-rases (underlined in Fig. 5). Three of the beef heart peptides(P3, P7, and P8) are similar to two different fumarasesegments, whereas one of the peptides (P9) is similar to onlyone cysteine-containing region. The similarity between P7

VOL. 171, 1989

on August 20, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

3500 BELL ET AL.

P5 VAVPSTIHeDHLIEAQLGGEK P7 DVGGIVLANAfCiGPeIGQWDRFUMA 1 SNKPFHYQAPFPLKKDDTEYYLLTSEHVSVSEFEGQE 37... 94 DIAAKGVLPT1CDTGTAIIVG 114..

* II*t

P3 VGLIGSCTNSSTEDMGR

P8 DFAPGKPLTI IKHPNGTQETILLNHTFNET+IEWFUA19DFgL* : : * * :.:*

FUMA 169 DLYVDGDEYKFLZIAKGGGSANKTYLYQETKALLTPGKLKNYLVEKM1GTAa PYHIAFVI 231.....

P9..GPGVDSIS* *

FUMA 297 IRVIRLPR)IGA_

T-GMATIjCfNMGAEIGAT**. 1* :* *

PVGMGVg ADRNIKAK 326..... 450 QGGSMIMLAKGNRSQQVTDA# KHGG 478wwIf v ,r

P3 VGLIGS CTNSSTEDMGR

P8 DFAPGKPL CIIKHPNGTQETILLNHTFNET+IEW P7 DVGGIVLANA GFHIGQWDR

FUMA 479 FYLGIIGGPAAVLAQGSIKSLEJ4VEYPELGMEAIWKIEVEDFPAFILVDDKGNDFFQQIQLT Jre K 547

FIG. 5. Sequence similarities between cysteinyl-containing segments of FUMA and cysteinyl-tryptic peptides of mammalian aconitases.Sequences resembling the reactive cysteinyl peptide of pig heart aconitase, IQLLCPLLNQFDK (9), are underlined. Similarities to cysteinylpeptides from beef heart aconitase, numbered according to the method of Plank and Howard (18), are shown (*) to signify identities and colonsindicating conservative substitutions scoring .0.1 in the MDM78 matrix (28). Cysteine residues are boxed, the hyphen in P9 denotes aninsertion which optimizes the similarity, and the plus sign in P8 denotes an unidentified residue.

and the C-terminal segment of fumarase is particularlyinteresting because two neighboring cysteine residues areconserved, and there is evidence relating disulfide formationin P7 to disruption of the Fe-S cluster of aconitase (22). Theexistence of such similarities is consistent with the view thatthe class I fumarases and aconitase are structurally relatedenzymes. Three of the aconitase peptides (P4, P5, and P10)exhibited significant homologies with unique fumarase seg-ments that lacked cysteine residues. In the case of P5, thereis a conserved histidine residue (Fig. 5), but this is replacedin FUMB. The peptide (P4) containing the reactive (butnonessential) cysteine residues of beef heart aconitaseshows no similarity with any of the cysteine-containingregions of the fumarase but is similar to a region containinga conserved tryptophan residue. The reactive cysteine pep-tide of the rat heart enzymes likewise shows no similarity toany of the cysteine peptides thus far defined for the beefheart enzyme.

Features of the unidentified genF product. The partialsequence of the gene located upstream of thefumB gene wasdesignated genF; the partial sequence of its product, GP-F,

co m

I091144420

~IH I

4430

11938

4440

is shown in Fig. 2. In an attempt to identify GP-F, itssequence was compared with those in a data base containingover 8,000 protein sequences, using the DAPSEARCH pro-gram (3, 4). This revealed an extremely strong similaritybetween GP-F and the product (GP-A) of another unidenti-fied E. coli gene (genA), which is located immediatelydownstream of the aspartase gene, aspA. An indication ofthe homology between GP-F and GP-A comes from theestimate that >1014 data bases of the same size would haveto be analyzed before a comparable match would occur bychance. No other significant homologies were detected. Therelative positions of fumB, genF, aspA, and genA in thelinkage and physical maps of E. coli are illustrated in Fig. 6.A partial sequence for genA was obtained previously in thislaboratory during the sequencing of the aspA and fumCgenes of E. coli K-12 (39). A complete sequence for the genAgene of E. coliW had likewise been obtained during analysisof the aspA gene of this strain (33), but there is now reasonto believe that this sequence is also incomplete. Initialcomparisons between GP-F and GP-A showed that thesequence similarity at the putative C terminus of GP-A

kgilI I

I I H H I HI9I0 942 944 min

4450 4460 kbFIG. 6. Relative positions of the unidentified genF and genA ger.es in the linkage and physical maps of the E. coli chromosome. The

locations and transcriptional polarities of relevant genes are shown, as are the positions of HindIll restriction sites (H). The scales identifypositions (minutes) in the linkage map (1) and nucleotide coordinates (kilobases) in the physical map (13).

FI11

J. BACTERIOL.

on August 20, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

FUMARASE B GENE OF E. COLI 3501

GP-A 1 MLVVELIIVLLAIFLGARLGGIGIGFAGGLGVLVLAAIGVKPGNIPFDVISIIMAVIAAI

GP-A

GP-F

I

61 SAMQVAGGLDYLVHQTEKLLRRNPKYITILAPIVTYFLTIFAGTGNISLATLPVIAEVAK.AS2D . ..22ML2ELRNK. .S.VP.CT 222. 22.L.. 22.2 .2A

(1) ....ASGGLDVMLQIAEKLLRRNPKYVSIVAPFVTCTLTILCGTGHVVYTILPIIYDVAI

GP-A 121 EQGVKPCRPLSTAVVSAQIAITASPISAAVVYMSSVM-----EGHGISYLHLLSVVIPSTF7 .22.2S.M222.2L.2 . ..L.. E.. . 2.I.

GP-F (57) KNNIRPERPMAASSIGAQMGIIASPVSVAVVSLVAMLGNVTFDGRHLEFLDLLAITIPST

GP-A

GP-F

176

(117)

GP-A 229

GP-F (175)

GP-A

GP-F

GP-A

GP-F

286

(231)

346

(291)

LLAVLVMSFLVTMLFNSK-LSDDPIYRKRL--EEGLVELRGEKQ-IEIKSGAKT---SVW$...*.... * .22 2 ..* . * 2. . * .2. ..*

LIGILAIGIF--SWFRGKDLDKDEEFQKFISVPENREYVYGDTATLLDKKLPKSNWLAMW

LFLLGVVGVVIYAIINS---PSMGLVEKPLMNTTNAILIIMLSVATLTTVICKVDTDNIL.2* 2 .22* . .2 2 2 222 2 .22 .2 .. 2 . g

IFL-GAIAVVALLGADSDLRPSFG--GKPL-SMVLVIQMFMLLTGALIIILTKTNPASIS

60

120

(56)

175

(116)

228

(174)

285

(230)

NSSTFKAGMSACICILGVAWLGDTFVSNNIDWIKDTAGEVIQGHPWLLAVIFFFASALLY 3452 . 2. 2. . . .. .. 2 22.. 22 2 . 2. .

KNEVFRSGMIAIVAVYGIAWMAETMFGAHMSEIQGVLGEMVKEYPWAYAIVLLLVSKFVN (290)

SQAATAKALIAMALALNVSR+DAVASFAAVSGLFILPTYPTLVAAVQMDDTGTTRIGKFV '405'222. 2..A..222. 2 222 A.2 G2 .222222. .22.2 2 .222.22.22

SQAAALAAIVPVALAIGVDPAYIVASAPACYGYYILPTYPSDLAAIQFDRSGTTHIGRFV (350)

GP-A '406' FNHPFFI.... '412'

GP-F (351) INHSFILPGLIGVSVSCVFGWIFAAMYGFL (380)FIG. 7. Alignment of partial amino acid sequences of the GP-A (genA) and GP-F (genF) proteins. The GP-A sequence was translated from

the genA region of E. coli W (33) after introduction of 2 bp at a stop codon (+) (see text). This extends the region of homology and maintainsthe alignment, but the sequences immediately adjacent to the frameshift (+) should be regarded as uncertain. The amino acid coordinates ofGP-F are bracketed because the sequence is preceded by an N-terminal segment of unknown length, and those at the C-terminal end of GP-Aare in inverted commas because of uncertainties about the effects of the frameshift: *, identical residues; :, conservative substitutions havingscores of .0.1 in the MDM78 matrix (28).

(residue 365) could be extended a further 46 residues byincorporating the whole of the published sequence previ-ously thought to lie outside the genA coding region in E. coliW (33). An insertion of 2 bp (or a deletion of 1 bp) close toor within the original stop codon would reestablish a readingframe having codon preferences that are typical of those ofE. coli genes. It would therefore appear that both of theinferred amino acid sequences are incomplete, GP-F at the Nterminus and GP-A at the C terminus.An alignment of the two partial sequences, in which the

similarity has been optimized by introducing several inser-tions or deletions, is shown in Fig. 7. Some 38% of 343equivalenced residues are identical, and the degree of simi-larity increases to 68% when conservative changes are

included. The GP-F and GP-A sequences have a very highproportion of hydrophobic residues, as is well illustrated bythe respective hydropathy profiles (Fig. 8). Both sequencescontain several hydrophobic segments resembling thosefound in integral membrane proteins. Furthermore, a-helicalsecondary structures were strongly predicted for some ofthese segments, which indicates that they are probablymembrane-spanning a-helices. If the homology between thetwo proteins is sustained through to comparable N-terminaland C-terminal extremities, it is likely that genF and genAeach encode a membrane protein of Mr 46,000 to 47,000.

Conclusion. This analysis of the fumB gene amply con-

firms that E. coli possesses two structurally analogous butdifferentially regulated fumarases of the relatively labiledimeric type (class I). Expression of the fumB gene isactivated under anaerobic conditions by FNR (37). Thiscontrasts with fumA expression, which has recently been

shown to be activated by the cyclic AMP-CRP complex andrepressed anaerobically by the arcA (dye)-encoded regulator(J. R. Guest and R. S. Buxton, unpublished observations).The kinetic properties of the two enzymes also indicate thatthe enzymes are specifically adapted for the dehydration ofmalate or the hydration of fumarate in order to performspecific roles in anaerobic metabolism associated with fuma-rate respiration (FUMB) or aerobic respiratory metabolisminvolving the citric acid cycle (FUMA). This situation ex-actly parallels that of fumarate reductase and succinatedehydrogenase, in which the corresponding frdABCD andsdhCDAB operons are subject to the same differential regu-lation by FNR in one case and the cyclic AMP-CRP complexplus the anaerobic repressor in the other, and the enzymecomplexes are specifically adapted to accept electrons frommenaquinione in fumarate respiration (fumarate reductase)or to donate electrons to ubiquinone in aerobic respiration(succinate dehydrogenase) (11, 17, 24, 35). The degree ofsequence conservation between the coregulated aerobic andanaerobic enzymes is particularly high for the fumarases(90% identity) compared with the flavoprotein and iron-sulfur protein subunits of the oxidoreductases (44 and 38%,respectively; 5, 36). The greater divergence of the latter mayreflect the importance of independent subunit interactionspecificities that may be needed for assembling the mem-brane-bound enzyme complexes.The function of the structurally distinct, stable, and tet-

rameric class II enzyme, FUMC, is not clear. Its kineticproperties more closely resemble those of the aerobic citricacid cycle enzyme, FUMA, but it is expressed under bothaerobic and anaerobic conditions, and the corresponding

VOL. 171, 1989

.

on August 20, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

3502 BELL ET AL.

(300) (380)

FIG. 8. Hydropathy profiles of the GP-A (genA) and GP-F(genF) partial sequences. Consecutive hydropathy averages were

plotted for each nine-residue span. The sequences were derived andaligned as described in the legend to Fig. 7.

fumC gene is capable of substituting for thefumA andfumBgenes (37, 40). The properties of the class I enzymespreviously indicated that they are related to aconitase andother hydrolyases in containing Fe-S clusters (17). This viewis strengthened by the similarities that have been detectedbetween the cysteine-containing regions of FUMA andFUMB and the mammalian aconitases. It also raises theintriguing question of whether there is a second class ofaconitases resembling FUMC and whether some organismscan express more than one class of aconitase.

It is noteworthy that the aspartase, asparaginase II, andfumarase B genes resemble the fumarate reductase operon inbeing positively regulated by FNR (12, 37). The enzymes arelikewise functionally related to the fumarate reduction sys-tem insofar as they are involved in the conversion ofaspaftate, asparagine, or malate to fumarate, the anaerobicelectron acceptor. It may also be significant that the aspA,fumB, and frdABCD genes are clustered in a small segmentof the linkage map and that the genes adjacent to aspA andfumB encode a pair of homologous and extremely hydro-phobic proteins. The corresponding gene orders differ,genF-fumB and aspA-genA, but in each case the intergenicregions are relatively small, and it is possible that the genepairs are coregulated or even cotranscribed. If so, the genAand genF products could perform analogous functions in, forexample, the membrane transport of aspartate and fumarateunder anaerobic conditions.

ACKNOWLEDGMENTS

We are indebted to S. A. Woods for help with the sequencingstrategy and oligonucleotide design.We are grateful for support from the University of Sheffield

Research Fund (P.J.B.) and the Science and Engineering ResearchCouncil (J.R.G.).

LITERATURE CITED1. Bachmann, B. J. 1983. Linkage map of Escherichia coli K-12,

edition 7. Microbiol. Rev. 47:180-230.2. Biggin, M. D., T. J. Gibson, and G. F. Hong. 1983. Buffer

gradient gels and 35S label as an aid to rapid DNA sequencedetermination. Proc. Natl. Acid. Sci. USA 80:3963-3965.

3. Collins, J. C., A. F. W. Coulson, and A. Fyall. 1988. Thesignificance of protein sequence similarities. CABIOS 4:67-71.

4. Coulson, A. F. W., J. C. Collins, and A. Lyall. 1987. Protein andnucleic acid sequence database searching: a suitable case forparallel processing. Comput. J. 30:420-424.

5. Darlison, M. G., and J. R. Guest. 1984. Nu&leotide sequenceencoding the iron-sulphurt; protein subunit of the succinatedehydrogenase of Escherichhfi coli. Biochem. J. 223:507-517.

6. Grosjean, H., and W. Fiers; 1982. Preferential codon usage inprokaryotic genes! the optimal codon-anticodon interactionenergy and the selective codon usage in efficiently expressedgenes. Gene 18:199-209.

7. Guest, J. R., J. S. Miles, R. E. Roberts, and S. A. Woods. 1985.The fumarase genes of Escherichia coli: location of the fumBgene and discovery of a new gene (fumC). J. Gen. Microbiol.131:2971-2984.

8. Guest, J. R., and R. E. Roberts. 1983. Cloning, mapping, andexpression of the fumarase gene of Escherichia coli K-12. J.Bacteriol. 153:588-596.

9. Hahm, K.-S., 0. Gawron, and D. Piszkiewicz. 1981. Amino acidsequence of a peptide containing an essential cysteine residue ofpig heart aconitase. Biochim. Biophys. Acta 667:457-461.

10. Henson, J. M., N. K. Blake, and L. E. Marek. 1987. Theisolation offumB mutants of Escherichia coli. J. Gen. Micro-biol. 133:2631-2638.

11. luchi, S., and E. C. C. Lin. 1988. arcA(dye), a global regulatorygene in Escherichia coli mediating repression of enzymes inaerobic pathways. Proc. Natl. Acad. Sci. USA 85:1888-1892.

12. Jerlstrom, P. G., J. Lui, and I. R. Beacham. 1987. Regulation ofEscherichia coli L-asparaginase II and L-asparatase by the fnrgene-product. FEMS Microbiol. Lett. 47:127-130.

13. Kohara, Y., K. Akiyama, and K. Isono. 1987. The physical mapof the whole E. coli chromosome: application of a new strategyfor rapid analysis and sorting of a large genomic library. Cell50:495-508.

14. Marinus, M. G., and N. R. Morris. 1973. Isolation of deoxyri-bonucleic acid methylase mutants of Escherichia coli K-12. J.Bacteriol. 114:1143-1150.

15. Messing, J. 1983. New M13 vectors for cloning. MethodsEnzymol. 101:20-78.

16. Miles, J. S., and J. R. Guest. 1984. Complete nucleotidesequence of the fumarase gene fumA, of Escherichia coli.Nucleic Acids Res. 12:3631-3642.

17. Miles, J. S., and J. R. Guest. 1987. Molecular genetic aspects ofthe citric acid cycle of Escherichia coli. Biochem. Soc. Symp.54:45-65.

18. Plank, D. W., and J. B. Howard. 1988. Identification of thereactive sulfhydryl and sequences of cysteinyl-tryptic peptidesfrom beef heart aconitase. J. Biol. Chem. 263:8184-8189.

19. Rosenberg, M., and D. Court. 1979. Regulatory sequencesinvolved in the promotion and termination ofRNA transciption.Annu. Rev. Genet. 13:319-353.

20. Sacchetini, J. C., M. W. Frazier, D. C. Chiara, L. J. Banaszak,and G. A. Grant. 1988. Amino acid sequence of porcine heartfumarase. Biochem. Biophys. Res. Commun. 153:435-440.

21. Sanger, F., A. R. Coulson, B. G. Barrell, A. J. H. Smith, andB. A. Roc. 1980. Cloning in single-stranded bacteriophage as anaid to rapid DNA sequencing. J. Mol. Biol. 143:161-178.

22. Schweiger, G., R. Dutscho, and W. Buckel. 1987. Purification of2-hydroxyglutarylCoA dehydratase from Acidaminococcus fer-mentans. An iron-sulfur protein. Eur. J. Biochem. 169:441-448.

23. Shaw, D. J., D. W. Rice, and J. R. Guest. 1983. Homologybetween CAP and FNR, a regulator of anaerobic respiration inEscherichia coli. J. Mol. Biol. 166:241-247.

24. Spiro, S., and J. R. Guest. 1987. Activation of the lac operon ofEscherichia coli by a mutant FNR protein. Mol. Microbiol.1:53-58.

GP-A100 200 300

3

2

1

'412'

,,,: I I

4-1

coCI0.L.0D

-1 Ii' r-2-

±3-2 | 1 1 4 1 r '1

2-

1-

-1

-2-

(1) (100) (200)GP-F

J. BACTERIOL.

1I I I -FF

on August 20, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

FUMARASE B GENE OF E. COLI 3503

25. Spiro, S., and J. R. Guest. 1987. Regulation and over-expressionof the fnr gene of Escherichia coli. J. Gen. Microbiol. 133:3279-3288.

26. Staden, R. 1979. A strategy of DNA sequencing employingcomputer programs. Nucleic Acids Res. 6:2601-2610.

27. Staden, R. 1980. A new computer method for the storage andmanipulation of DNA gel reading data. Nucleic Acids. Res.8:3673-3694.

28. Staden, R. 1982. An interactive graphics program for comparingand aligning nucleic acid and amino acid sequences. NucleicAcids Res. 10:2951-2961.

29. Staden, R. 1984. Graphic methods to determine the function ofnucleic acid sequences. A summary of ANALYSEQ options.Nucleic Acids Res. 12:505-519.

30. Staden, R., and A. D. McLachlan. 1982. Codon preference andits use in identifying protein coding regions in long DNAsequences. Nucleic Acids Res. 10:141-156.

31. Stormo, G. D., T. D. Schneider, L. M. Gold, and A. Ehrenfeucht.1982. Use of the 'Perceptron' algorithm to distinguish transla-tional initiation sites in E. coli. Nucleic Acids Res. 10:2997-3011.

32. Tabor, S., and C. C. Richardson. 1978. DNA sequence analysiswith a modified bacteriophage T7 DNA polymerase. Proc. Natl.Acad. Sci. USA 84:4767-4771.

33. Takagi, J. S., N. Ida, M. Tokushige, H. Sakamoto, and Y.Shimura. 1985. Cloning and nucleotide sequence of the aspar-

tase gene of Escherichia coli W. Nucleic Acids Res. 13:2063-2074.

34. Tinoco, I., Jr., P. N. Borer, B. Dengler, M. D. Levine, 0. C.Uhlenbeck, D. M. Crothers, and J. Gralla. 1973. Improvedestimation of secondary structure in ribonucleic acids. Nature(London) New Biol. 246:40-41.

35. Wilde, R. J., and J. R. Guest. 1986. Transcript analysis of the

citrate synthase and succinate dehydrogenase genes of Esche-richia coli K12. J. Gen. Microbiol. 132:3239-3251.

36. Wood, D., M. G. Darlison, R. J. Wilde, and J. R. Guest. 1984.Nucleotide sequence encoding the flavoprotein and hydropho-bic subunits of the succinate dehydrogenase of Escherichia coli.Biochem. J. 222:519-534.

37. Woods, S. A., and J. R. Guest. 1987. Differential roles of theEscherichia coli fumarases and fnr-dependent expression offumarase B and aspartase. FEMS Microbiol. Lett. 48:219-224.

38. Woods, S. A., J. S. Miles, and J. R. Guest. 1988. Sequencehomologies between argininosuccinase, aspartase and fuma-rase: a family of structurally-related enzymes. FEMS Micro-biol. Lett. 51:181-186.

39. Woods, S. A., J. S. Miles, R. E. Roberts, and J. R. Guest. 1986.Structural and functional relationships between fumarase andaspartase. Nucleotide sequences of the fumarase (fumC) andaspartase (aspA) genes of Escherichia coli K12. Biochem. J.237:547-557.

40. Woods, S. A., S. D. Schwartzbach, and J. R. Guest. 1988. Twobiochemically-distinct classes of fumarase in Escherichia coli.Biochim. Biophys. Acta 954:14-26.

41. Yanisch-Perron, C., J. Vieira, and J. Messing. 1985. ImprovedM13 phage cloning vectors and host strains: nucleotide se-quences of the M13 mpl8 and pUC19 vectors. Gene 33:103-119.

42. Yazyu, H., S. Shiota-Niiya, T. Shimamoto, H. Kamazawa, M.Futai, and T. Tsuchiya. 1984. Nucleotide sequence of the melBgene and characteristics of deduced amino acid sequence of themelibiose carrier in Escherichia coli. J. Biol. Chem. 259:4320-4326.

43. Yumoto, N., and M. Tokushige. 1988. Characterization of mul-tiple fumarase proteins in Escherichia coli. Biochem. Biophys.Res. Commun. 153:1236-1243.

VOL. 171, 1989

on August 20, 2018 by guest

http://jb.asm.org/

Dow

nloaded from