Biochimica et Biophysica Acta 1472 (1999) 376^384

www.elsevier.com/locate/bba

The gene for 2-phosphoglycolate phosphatase (gph) in Escherichia coli is

located in the same operon as dam and at least ¢ve other diverse genes
Anita Lyngstadaas *, Anders LÖbner-Olesen, Ellen Grelland, Erik Boye
Department of Cell Biology, Institute of Cancer Research, Montebello, N-0310 Oslo, Norway

Received 6 May 1999; received in revised form 2 August 1999; accepted 5 August 1999

Abstract

Downstream of the dam gene in the Escherichia coli genome the following three genes are located: first rpe, then a gene
encoding a 27 kDa protein and finally trpS. Here we present evidence that the 27 kDa protein has 2-phosphoglycolate
phosphatase activity, and we name the gene gph. Phosphoglycolate phosphatase is needed in autotrophic organisms
performing the Calvin^Benson^Bassham (CBB) reductive pentose^phosphate cycle. E. coli is not capable of autotrophic
growth and probably utilizes Gph activity for other function(s) than in the CBB cycle. We found no physiological effect of
deleting gph and its function in E. coli remains unclear. The use of fusion plasmids, where lacZ was inserted into gph and
trpS, and deletion derivatives of these fusion plasmids, showed that rpe, gph and trpS are all members of the dam-containing
operon. A novel promoter was identified in the distal part of the dam gene. The operon, which contains aroK, aroB, urf74.3,
dam, rpe, gph, and trpS, can be termed a superoperon, since it consists of (at least) seven apparently unrelated genes which are
under complex regulatory control. ß 1999 Elsevier Science B.V. All rights reserved.

Keywords: dam containing operon; Phosphoglycolate phosphatase; Superoperon; (Escherichia coli)

1. Introduction mapped in this region [4,10]. The major promoter

The Escherichia coli dam gene, encoding DNA ad-
enine methyltransferase [1,2], is located at the 3.51
Mb position of the genome [3]. It is part of an oper-
on containing at least four genes [4], in order: aroK
encoding shikimic acid kinase I [5,6], aroB encoding
3-hydroquinate synthase [7], urf74.3 encoding a pro-
tein of unknown function [8,9] and dam (Fig. 1). Five
transcription initiation sites, termed P1^P5, and two
transcription termination sites (T1, T2) have been
* Corresponding author. Present address: Norwegian Centre
for Health Technology Assessment, SINTEF Unimed, Pb 124
Blindern, 0314 Oslo, Norway. Fax: +47-22067979;
E-mail: anita.lyngstadaas@unimed.sintef.no
for the operon, P2, is located 3.5 kb upstream of
the dam gene.
The following three genes are found downstream
of dam (in order): rpe, a gene encoding a 27 kDa
protein, and trpS. The rpe gene encodes ribulose-5-
phosphate 3-epimerase activity [11] and trpS encodes
the tryptophanyl-tRNA synthetase [12]. The rpe gene
is separated from dam by 17 nucleotides (nt) [9]. The
short intergenic region between rpe and the 27kDa
gene (3 nt) and the 8 nt overlap between the open
reading frames of the 27kDa gene and trpS suggest a
transcriptional unit that consists of at least the rpe,
27kDa and trpS genes [9]. In addition, we have pre-
sented evidence for translational coupling between
rpe and the 27kDa gene [9].

0304-4165 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved.
PII: S 0 3 0 4 - 4 1 6 5 ( 9 9 ) 0 0 1 4 6 - 4

BBAGEN 24895 1-10-99

 A. Lyngstadaas et al. / Biochimica et Biophysica Acta 1472 (1999) 376^384
Based on amino acid sequence analysis, we pro-
posed that the 27kDa gene encodes 2-phosphoglyco-
late phosphatase (EC 3.1.3.18), an enzyme that con-
verts 2-phosphoglycolate (PG) to glycolate [9]. The
amino acid sequence of the 27 kDa protein is similar
to that of phosphoglycolate phosphatase from auto-
trophic organisms like Ralstonia eutropha (previously
Alcaligenes eutrophus) and Rhodobacter sphaeriodes
[13,14]. The principal route of CO2 assimilation in
autotrophic organisms, the Calvin^Benson^Bassham
(CBB) reductive pentose^phosphate cycle, produces
PG as a side-product, and phosphoglycolate phos-
phatase activity avoids PG accumulation and allows
it to be recycled into the CBB cycle. Most autotro-
phic organisms have their CBB enzymes, including
ribulose-5-phosphate 3-epimerase, encoded within
cbb clusters [15,16].
Here we have studied the function of the 27 kDa
protein. Enzymatic assays show that the protein has
2-phosphoglycolate phosphatase activity, and we
name the gene gph. Removing or increasing gph ex-
pression has no physiological e¡ect under the condi-
tions tested, and the function of Gph in E. coli re-
mains unclear. Furthermore, we have analyzed the
expression pattern of the genes surrounding gph.
We ¢nd that rpe, gph and trpS belong to the dam-
containing operon, which therefore contains at least
seven genes with diverse functions.
2. Materials and methods
2.1. Growth media
Cells were grown in either Luria Broth (LB) [17] or
in AB minimal medium [18] supplemented with 1 Wg/
ml thiamine and 0.2% glucose. Casamino acids
(CAA; Bacto Difco) were added to a ¢nal concen-
tration of 0.5%. Leucine was supplemented at 40 Wg/
ml. Selective growth of strains was achieved by addi-
tion of ampicillin or kanamycin at ¢nal concentra-
tions of 25 Wg/ml and 30 Wg/ml, respectively.
2.2. Bacterial strains and plasmids
The E. coli K-12 derived strains and plasmids used
are shown in Table 1. Plasmid pANL15 is the
pUHE21^2 expression vector [19] containing gph
BBAGEN 24895 1-10-99
377
under the control of an inducible promoter [9]. Plas-
mid pJEL109 (J.E.L. Larsen, unpublished; [4]) is an
R1-derived vector with a copy number of one to two
per host chromosome [20]. It contains the R1 origin
of replication, the bla gene of Tn3, unique cloning
sites for the restriction enzymes BamHI, ClaI and
HindIII, and a unique AseI site in the bla gene.
2.3. Preparation of cell extracts and assay of
2-phosphoglycolate phosphatase activity
Strains were grown in LB supplemented with glu-
cose to an optical density (OD at 600 nm) of 0.5,
harvested by centrifugation, resuspended in ice-cold
bu¡er (20 mM Tris^HCl (pH 7.6) containing 10 mM
MgCl2 and 1 mM dithioerythritol), and disrupted by
sonication on ice. Protein concentration was meas-
ured as described [21]. Phosphoglycolate phosphatase
activity was assayed by colorimetric measurements of
free phosphate production (OD at 820 nm) [13]. Ex-
tracts from strain ANL99 were prepared after
growth in the presence of 1 mM IPTG (isopropyl-
L-D-galactopyranoside), to allow induction of the
cloned gph gene.
2.4. Growth analysis
For growth rate analysis, cells in a single colony
were picked from a fresh LB plate and propagated in
the medium of interest. Growth was analyzed by
measuring OD at 600 nm (LB) or 450 nm (minimal
medium) with appropriate dilutions into the same
medium if necessary. The cultures were grown in
shaking waterbaths at 37³C. All growth rate experi-
ments were performed at least twice.
2.5. L-Galactosidase assay
Strain LJ24 containing the di¡erent lacZ fusion
plasmids were grown in AB minimal medium supple-
mented with CAA and glucose and the L-galactosidase
activity measured. Cells were permeabilized by sonica-
tion on ice and L-galactosidase units calculated as de-
scribed [17]. Enzyme levels of each strain were deter-
mined as an average of ¢ve samples taken at di¡erent
optical densities of the culture. The values presented in
Table 3 are mean values of at least two independent
measurements with standard deviation 6 10%.
378 A. Lyngstadaas et al. / Biochimica et Biophysica Acta 1472 (1999) 376^384

2.6. S1 nuclease mapping
RNA was puri¢ed from exponentially growing
cells as described [22]. S1 nuclease mapping was per-
formed essentially as described before [23], with
some modi¢cations. The RNA probe covering the
HpaI^HincII region was synthesized from a plasmid
with this region cloned into the EcoRV and SalI
sites of a Bluescript vector. T7 polymerase was
used for transcription of the linearized fragment in
the presence of K-35 S-labeled UTP. The probe prep-
aration was treated with DNaseI and hybridized to
total RNA in the presence of yeast tRNA before
treatment with S1 nuclease. Total probe length was
510 nt, covering the chromosomal fragment (456 nt)
plus 37 and 26 nucleotides of vector sequence (see
Fig. 2B). The protected fragments were run out on a
polyacrylamide gel and the radioactive bands were
detected in a Phosphoimager. Fragment lengths were
determined by comparing with a parallel sequencing
reaction.
2.7. Construction of gph: :lacZ and of trpS: :lacZ
fusion plasmids
Plasmid pALO160 [4] is pJEL109 carrying the 8 kb
BglII^BglII fragment of interest (Fig. 1). The
gph: :lacZ transcriptional fusion plasmid pEBO7
was constructed by inserting the promoterless lacZ
gene, carried on a 3 kb BamHI^DraI fragment from
plasmid pTL25, between the two closely spaced StuI
sites of plasmid pALO160. Insertion of the blunted
BamHI end of the lacZ fragment into the StuI site of
pALO160 created a BamHI restriction site in the
resulting plasmid pEBO7, a site that was used in
the cloning described below. Another gph: :lacZ
transcriptional fusion plasmid, pEBO6, is similar to
pEBO7, but the 3 kb HindIII^DraI lacZ fragment
from pTL25 was inserted into the StuI site(s) of
pALO160 after blunting of the HindIII site. Control
experiments showed that the lacZ reporter genes,
whether we used the BamHI^DraI or the HindIII^
DraI fragment, were equivalent. The trpS: :lacZ
transcriptional fusion plasmid pANL51 was con-

Table 1
Bacterial strains and plasmids
Strain/plasmid Relevant characteristics Source or
reference
Strains
LJ24 leu-6 v lac [47]
ANL9 LJ24 gph: :Km [9]
ANL99 LJ24 containing pANL15, Gph overproducer This work
AB1157 wt [48]
MG3819 AB1157 dam16: :Km [49]

Plasmids
pJEL109 R1-derived vector, cloning vector for L-galactosidase assay [4]
pTL25 Contains lacZ on a 3 kb BamHI^DraI (B^D) or HindIII^DraI (H^D) fragment [50]
pALO160 pJEL109 with aroK^trpS (BglII^BglII) in the BamHI site [4]
pALO163 pALO160 with lacZ (B^D) in the BamHI site; dam: :lacZ fusion plasmid [4]
pALO163#87 Bal31 deletion derivative of pALO163 missing bp 1740^2929 (Fig. 1) This work
pALO226 pEBO6 with Bal31 deletion (bp 1740^2929, v P2); gph: :lacZ fusion plasmid This work
pEBO6 pALO160 with lacZ (H^D) in StuI site; gph: :lacZ fusion plasmid This work
pEBO7 pALO160 with lacZ (B^D) in StuI site; gph: :lacZ fusion plasmid This work
pEBO8 pEBO7 with AseI^EcoRI (vP1^P3) deletion; gph: :lacZ fusion plasmid This work
pANL15 pUHE21-2 with gph (AviII^AluI) in the BamHI site; gph expression construct [9]
pANL51 pALO160 with lacZ (B^D) in the Bsu36I site; trpS: :lacZ fusion plasmid This work
pANL53 pANL51 with AseI^EcoRI deletion (vP1^P3); trpS: :lacZ fusion plasmid This work
pANL54 pANL51 with AseI^BamHI deletion (v P1^P5); trpS: :lacZ fusion plasmid This work
pANL55 pEBO7 with AseI^BamHI deletion (v P1^P5); gph: :lacZ fusion plasmid This work
pANL59 pEBO7 with Bal31 deletion (bp 1740^2929, v P2); gph: :lacZ fusion plasmid This work
pANL60 pANL51 with Bal31 deletion (bp 1740^2929, v P2); trpS: :lacZ fusion plasmid This work

BBAGEN 24895 1-10-99

 A. Lyngstadaas et al. / Biochimica et Biophysica Acta 1472 (1999) 376^384
Table 2
Gph activity in cell extracts from di¡erent strains
Strain Gph activitya (U/mg protein)
LJ24 (wt) 0.03
ANL9 (gph: :km) 6 0.001
ANL99 (pUHE^gph) 8
a
One unit is the amount of cell extract which produces 1 Wmol
of product in 1 min.
structed by inserting the BamHI^DraI fragment of
lacZ from pTL25 into the unique Bsu36I site of
pALO160, after blunting of cohesive ends. All blunt-
ing reactions were performed by Klenow enzyme.
2.8. Deletion derivatives of the gph: :lacZ fusion
plasmid pEBO7 and pEBO6
For constructing pANL55, plasmid pANL54 (de-
scribed below), lacking the 4.5 kb AseI^BamHI frag-
ment which contains promoters P1 through P5 (Fig.
1) was used as follows: the HindIII^EagI fragment
of pANL54, from which P1^P5 has been removed in
pANL54, and the EagI^HindIII gph: :lacZ-contain-
ing fragment of pEBO7 were ligated.
For constructing plasmid pEBO8, pEBO7 was di-
gested with AseI and EcoRI, resulting in three frag-
ments. The 2.8 kb AseI^EcoRI P1^P3-containing
chromosomal fragment (Fig. 1) was removed after
gel electrophoresis and the two remaining fragments
were ¢rst religated at their AseI ends followed by
treatment with Klenow enzyme and ligation.
A library of deletion derivatives of the dam: :lacZ
fusion plasmid pALO163 [4], was used for construct-
ing pANL59. Plasmid pALO163 Bal31#81 from this
library is missing basepairs no 1740 to 2929, which
are replaced by a BamHI linker, thereby removing
the 310 region of P2, whereas P1 is intact (see Fig.
1). Digestion of pALO163 Bal31#81 and pEBO6
with AseI and EcoRI results in three fragments of
each plasmid. By putting the 1.6 kb AseI^EcoRI
fragment of pALO163 Bal31#81 into pEBO6 instead
of its 2.8 kb AseI^EcoRI fragment, we made
pALO226. Plasmid pANL59 was constructed by
the same strategy, by replacing the 2.8 kb AseI^
EcoRI fragment of pEBO7 with the 1.6 kb AseI^
EcoRI fragment of pALO226.
BBAGEN 24895 1-10-99
379
Fig. 1. Map of the dam-containing operon. Top scale in kilo-
bases (kb). Previously reported promoters and terminators in
the operon are numbered P1^P5 and T1^T2, respectively. Puta-
tive promoter, as reported in this study, is numbered P6. Only
restriction sites used in this study are shown (abbreviations: Bg,
BglI; As, AseI; E, EcoRI; B, BamHI; Ea, EagI; Hp, HpaI;
Hi, HincII ; A, AluI; Av, AviII; St, StuI; Bs, Bsu36I; H, Hin-
dIII). The two Ba entries represent the endpoints of a deletion
created by the use of Bal31 nuclease. pJEL109 is the cloning
vector. Restriction sites for insertions of lacZ are marked X.
Regions of the operon deleted in certain constructs are marked
with dotted lines.
2.9. Deletion derivatives of the trpS: :lacZ fusion
plasmid pANL51
For constructing a P1^P5 deletion version of the
trpS: :lacZ fusion plasmid, pANL51 was digested
with AseI and BamHI, resulting in three fragments.
The 4.5 kb P1^P5-containing chromosomal fragment
was removed and the remaining fragments were ¢rst
religated at their AseI ends followed by treatment
with Klenow enzyme and ligation to create pANL54.
The P1^P3 deletion of pANL51, pANL53, was
constructed by digesting pANL51 with AseI and
EcoRI, and the fragments were treated as described
for pEBO8 (see above).
Plasmid pANL60 (containing the 1.2 kb Bal31 de-
letion) was constructed by replacing the 2.8 kb AseI^
EcoRI fragment of pANL51 with the 1.6 kb AseI^
EcoRI fragment of pALO226, using the same strat-
egy as for constructing pANL59 (see above).
380 A. Lyngstadaas et al. / Biochimica et Biophysica Acta 1472 (1999) 376^384
3. Results
3.1. The 27 kDa protein has 2-phosphoglycolate
phosphatase activity
Enzymatic assays were performed to test the hy-
pothesis that the 27kDa gene encodes phosphoglyco-
late phosphatase. We prepared crude cell extracts
from the wild-type (LJ24), the 27kDa: :Km disrup-
tion mutant (ANL9) and 27 kDa overproducer
(ANL99) cells and assayed them for 2-phosphogly-
colate phosphatase activity. The assay revealed a low
activity in wild-type cells and no detectable activity
in the disruption mutant (Table 2). In extracts from
the overproducer, the enzymatic activity was two or-
ders of magnitude higher than that found in wild-
type extracts. We conclude that the 27 kDa protein
has phosphoglycolate phosphatase activity and that
this is the only detectable phosphoglycolate phospha-
tase activity in E. coli. Since the gene notation pgp is
already in use for E. coli, we suggest the name gph
for the gene (phosphoglycolate phosphatase). It has
earlier been reported that E. coli does not contain
phosphoglycolate phosphatase activity [13]. The level
of activity reported in the present work is low and its
unambiguous detection was dependent upon con-
struction of the gph disruption mutant and the
Gph overproducer.
3.2. Normal growth of the gph mutant
In contrast to a preliminary report on the growth
of the gph mutant [9], we report here that it has the
same doubling time (23 min) as wild-type cells in LB
supplemented with glucose. Also, no di¡erence was
found in minimal medium with glucose (doubling
time 65 min), or with glucose and CAA (30 min).
Table 3
L-Galactosidase activity (units) of lacZ fusion plasmids and
their deletion derivatives
No deletions vP1^P5 vP1^P3
dam^lacZ 350 ^ ND
gph^lacZ 1000 125 250
trpS^lacZ 1100 350 450
BBAGEN 24895 1-10-99
3.3. Transcriptional activity downstream of dam
To analyze transcription downstream of dam dif-
ferent lacZ transcriptional fusion plasmids were
made. The lacZ gene was inserted into either the
BamHI site in dam, the StuI site in gph or the
Bsu36I site in trpS (Fig. 1). The L-galactosidase ac-
tivities in cells containing the gph: :lacZ fusion and
the trpS: :lacZ fusion were 1000 U and 1100 U, re-
spectively (Table 3), i.e., three times higher than the
activity measured for the dam: :lacZ fusion (350 U),
suggesting promoter activity downstream of dam.
3.4. dam transcription continues through rpe, gph and
trpS
To further investigate the transcriptional activity
in the region downstream of dam the fusion plasmids
with lacZ inserted into gph and trpS were used to
construct a series of deletion derivatives lacking the
di¡erent promoters upstream of dam (P1^P5; see
Fig. 1). The activity produced by the gph: :lacZ fu-
sion plasmid lacking P1 through P5 was 125 U com-
pared to 1000 U for the parent fusion plasmid (Table
3). Also, deletion of the P1^P3 region and of the P2
region from the gph: :lacZ fusion plasmid yielded
dramatically reduced L-galactosidase activities, 250
U and 200 U, respectively. Thus, expression of gph
is reduced by deletion of promoters upstream of dam.
Furthermore, gph expression is mostly dependent
upon P2, the major promoter for the dam operon.
We conclude that gph, and therefore rpe, are mem-
bers of the same operon as dam.
Removal of the P1^P5, P1^P3 or P2 regions from
the trpS: :lacZ fusion plasmid resulted in L-galacto-
sidase activities of 350 U, 450 U and 220 U, respec-
tively (Table 3) as compared to the 1100 U meas-
ured for the parent fusion plasmid. Therefore, the
upstream promoters, and in particular P2, contrib-
ute signi¢cantly to trpS expression as well. We con-
clude that trpS is also part of the dam-containing
operon, which thus contains (at least) the seven
genes, in order, aroK, aroB, urf74.3, dam, rpe, gph
and trpS.
vP2
ND 3.5. A distal promoter in the dam gene
200
220
The above L-galactosidase measurements locate
 A. Lyngstadaas et al. / Biochimica et Biophysica Acta 1472 (1999) 376^384
Fig. 2. Locating the P6 promoter by S1 nuclease protection.
(A) Protected probe visualized by a Phosphoimager. Lane 1,
undigested probe; lane 2, digested probe, no E. coli RNA
added; lane 3, RNA from the wild-type strain AB1157; lane 4,
RNA from the dam16: :Km mutant GM3819. Fragment lengths
are given in nucleotides. (B) Schematic diagram of the labelled
RNA probe (bottom) aligned with the chromosomal fragment
(top) and the 390 bp protected fragment. (C) The nucleotide se-
quence of the P6 promoter. Shown are the 310 and 335 ele-
ments (boxed). The leftmost A is located at position 5738 of
Fig. 1.
one or more promoter(s) in the region between the
dam and gph insertion sites, i.e., within the 1.6 kb
BamHI^StuI fragment (Fig. 1). In an earlier study
[9], we presented evidence for a promoter in the dis-
tal 140 nt of the dam gene. S1 nuclease assays re-
vealed the location of this promoter. An RNA probe
(Fig. 2B) spanning the HpaI^HincII region (Fig. 2A,
lane 1) was either protected in the whole chromoso-
mal region by total E. coli RNA (Fig. 2A, lanes 3
and 4) or reduced in size by 66 nt (Fig. 2A, lanes 3
and 4). No protected fragment could be detected in
the absence of E. coli RNA (Fig. 2A, lane 2). The
shortest protected probe was present in relatively
higher amount in a dam16: :Km deletion/insertion
mutant than in the otherwise genetically identical
BBAGEN 24895 1-10-99
381
wild type (Fig. 2A, lane 4). The mutant lacks the
chromosomal EcoRV^HpaI fragment (Fig. 1) and
therefore retains the probed region. Separate experi-
ments, involving end-labeled DNA probes, demon-
strated that there are no transcription termination
sites in the region (data not shown), showing that
the shortest fragment (Fig. 2) arises because of a
promoter located close to the HpaI site. The size of
the fragment is indeed consistent with the location of
a sequence very similar to that of c70 -dependent pro-
moters, with 310 and the 335 elements resembling
consensus (Fig. 2C). We term this promoter P6 (Fig.
1). The fragment between the 456 nt and 390 nt frag-
ments (Fig. 2A, lanes 3 and 4) suggests yet another,
but weaker, promoter in the region, but no sequence
close to consensus could be found at the appropriate
location.
4. Discussion
4.1. A gph gene in E. coli
We have shown that the gene downstream of rpe
in the E. coli genome encodes phosphoglycolate
phosphatase. The two E. coli genes rpe and gph
both have high similarities to genes encoded in cbb
clusters encoding enzymes of the Calvin^Benson^
Bassham (CBB) reductive pentose^phosphate cycle,
which represents the main biochemical pathway for
CO2 assimilation in autotrophic organisms. Presum-
ably, the two genes have been moved together into
their present location from a cluster of CBB genes.
The same may be true for the close relative, Haemo-
philus in£uenzae, where rpe and gph homologues are
colocalized as in E. coli [24]. Ribulose-5-phosphate 3-
epimerase of autotrophic organisms functions in the
pentose^phosphate pathway (PPP) as well as in the
CBB cycle. Recently, the E. coli rpe was demon-
strated to function in the PPP [11]. The only known
function of phosphoglycolate phosphatase, however,
is associated with the CBB cycle of autotrophic or-
ganisms.
4.2. Possible physiological functions of Gph in E. coli
The function of Gph activity in E. coli is unclear.
The metabolic pathway(s) utilizing Gph and the ex-
382 A. Lyngstadaas et al. / Biochimica et Biophysica Acta 1472 (1999) 376^384
tent of intracellular phosphoglycolate (PG) produc-
tion is unknown. We ¢nd a low but detectable Gph
activity in wild-type cells. No physiological e¡ect is
observed by disrupting the gph gene (this work) or by
overexpressing gph [9].
We shall consider the possibility that Gph func-
tions in a so far unidenti¢ed CBB or CBB-like cycle
of E. coli. There are two enzymes of the CBB cycle
that are not identi¢ed in E. coli: ribulose-5-phos-
phate kinase (Prk) and the central CBB enzyme, ri-
bulose-biphosphate carboxylase/oxygenase (Rubis-
CO). The E. coli genome indeed contains a gene
encoding a putative Prk homolog [25]. This enzyme
might produce ribulose-diphosphate (Rul-1,5-P2 ),
which is the substrate for RubisCO. RubisCO acts
as a carboxylase producing 3-phosphoglycerate from
Rul-1,5-P2 . At high oxygen concentrations (and low
CO2 ) RubisCO can act as an oxygenase producing
PG from Rul-1,5-P2 . In E. coli this would give a role
to Gph, namely to convert PG to glycolate, which,
when further metabolized, leads to the glycolytic
pathway and TCA cycle intermediates [26,27]. How-
ever, there is no evidence for a RubisCO-like enzyme
in E. coli. Recently, genes and enzymes thought to be
speci¢cally involved in the CBB cycle have been
found in other bacterial species incapable of autotro-
phic growth. Examples include a RubisCO homolog
in Bacillus subtilis [28], a Prk homolog in Erwinia
chrysanthemi [29] and RubisCO activity in archae-
bacteria like Haloferax mediterranei [30,25]. Thus,
there is a possibility that these species, including E.
coli, perform a complete or modi¢ed CBB cycle. Al-
ternatively, the Gph protein has a novel function
unrelated to the CBB cycle.
4.3. An operon containing the genes
aroK-aroB-urf74.3-dam-rpe-gph-trpS
An operon consists of genes that are cotranscribed
into the same mRNA, including adjacent cis-acting
sequences required for transcription of the genes
[31]. As described here, expression of rpe, gph and
trpS is mostly dependent upon P2, the major pro-
moter for the aroK-aroB-urf74.3-dam cluster (Fig. 1).
Thus, we have provided evidence that the dam-con-
taining operon consists of (at least) the seven genes
aroK, aroB, urf74.3, dam, rpe, gph and trpS, which
spans 7 kb.
BBAGEN 24895 1-10-99
4.4. The operon encodes apparently unrelated gene
products
A common denominator for the functions of the
seven genes in the dam-containing operon is not evi-
dent. AroB and AroK participate in early reactions
in the common pathway of aromatic amino acid bio-
synthesis [32]. Interestingly, loss of AroK confers
mecillinam resistance [33], indicating that AroK has
a second activity, possibly related to cell division.
Urf74.3 may be involved in cell division, but its func-
tion is dispensable [9]. Dam methylation is involved
in several important processes related to the cell
cycle, such as DNA replication [34], mismatch repair
(reviewed in [35]), gene expression [36,37], and trans-
position frequency [38]. Rpe acts in the interconver-
sion between ribulose-5-phosphate and xylulose-5-
phosphate in the non-oxidative branch of the PPP
[11]. Surprisingly, the rpe gene product was recently
reported to be involved in the control of chromo-
some replication in E. coli [39], suggesting that Rpe
has a second activity related to cell cycle control. The
function of Gph is unclear, as discussed above. TrpS
catalyzes the formation of tryptophanyl-tRNA, acti-
vating tryptophan for incorporating into proteins. A
collection of genes with so diverse functions into one
operon is highly unusual for E. coli.
4.5. Complex regulatory control of the operon
There are ¢ve promoters in the operon upstream
of dam and we have identi¢ed a sixth promoter close
to the end of dam. In addition, we have preliminary
evidence that there are two more promoters located
between dam and trpS. The P2 promoter is the stron-
gest and ensures a basal level of transcription for
most of the genes in the operon, whereas the other
promoters modulate or ¢ne-tune expression of an
individual gene or groups of genes. The presence of
multiple promoters suggests that transcriptional reg-
ulation is complex.
P2 is growth rate regulated [40,41], i.e. the activity
increases with growth rate, suggesting that one or
more of the downstream genes require growth rate
regulation. As suggested earlier [40] this may be the
dam gene, but the present work adds rpe, gph and
trpS as candidates for growth rate regulation. Some
central metabolic pathway genes are known to be
 A. Lyngstadaas et al. / Biochimica et Biophysica Acta 1472 (1999) 376^384
growth rate regulated [42] and this may also be the
case for rpe. Interestingly, the P6 promoter has an
AT-rich region immediately upstream of the 335
region that might function as a UP element [43], an
element characteristic of some growth rate regulated
promoters [44]. There is evidence that trpS is growth
rate regulated, but this regulation is mediated by
sequences downstream of dam [45]. In fact, a com-
mon denominator for the genes in the operon may be
that they require and/or prefer growth rate regula-
tion.
4.6. A superoperon
The dam-containing operon ful¢ls the require-
ments for being termed a superoperon [46], since:
(i) the operon consists of several genes spanning a
long region, (ii) the operon encodes apparently un-
related gene products and (iii) expression of the
genes in the operon is under complex regulatory
control.
4.7. Use of the lacZ reporter gene: physiological
considerations
Intentionally, we have not used a chromosome-
based reporter gene in the present experiments, since
some of the mutations/deletions studied a¡ect the
carbon £ow, and therefore the growth rates, of the
strains [11]. Instead, we employed a low copy num-
ber vector containing the reporter gene, so that in all
our host cells the full complement of genetic infor-
mation was always present. This allowed us to per-
form the reporter gene assays under conditions where
the host cells, in all cases, grew with the same dou-
bling time. We consider it extremely unlikely that the
plasmid mutations under these conditions should af-
fect plasmid copy number and thereby in£uence the
reporter gene measurements, although this was not
measured directly.
Acknowledgements
This work was supported by the Norwegian Can-
cer Society. We are also grateful for support from
Anne-Brit KolstÖ, Institute of Pharmacy, University
of Oslo.
BBAGEN 24895 1-10-99
383
References
[1] S. Lacks, B. Greenberg, J. Mol. Biol. 114 (1977) 153^168.
[2] S.M. Hattman, J.E. Brooks, M. Masurekar, J. Mol. Biol.
126 (1978) 367^380.
[3] F.R. Blattner, G. Plunkett III, C.A. Bloch, N.T. Perna, V.
Burland, M. Riley, J. Collado-Vides, J.D. Glasner, C.K.
Rode, G.F. Mayhew, J. Gregor, N.W. Davis, H.A. Kirkpa-
trick, M.A. Goeden, D.J. Rose, B. Mau, Y. Shao, Science
277 (1997) 1453^1462.
[4] A. LÖbner-Olesen, E. Boye, M.G. Marinus, Mol. Microbiol.
6 (1992) 1841^1851.
[5] A. LÖbner-Olesen, M.G. Marinus, J. Bacteriol. 174 (1992)
525^529.
[6] M.J. Whipp, A.J. Pittard, J. Bacteriol. 177 (1995) 1627^1629.
[7] G. Millar, J.R. Coggins, FEBS Lett. 200 (1986) 11^17.
[8] P. Jonczyk, R. Hines, D.W. Smith, Mol. Gen. Genet. 217
(1989) 85^96.
[9] A. Lyngstadaas, A. LÖbner-Olesen, E. Boye, Mol. Gen.
Genet. 257 (1995) 546^554.
[10] T.-H. Wu, E. Grelland, E. Boye, M.G. Marinus, Biochim.
Biophys. Acta 1131 (1992) 47^52.
[11] A. Lyngstadaas, G.A. Sprenger, E. Boye, Biochim. Biophys.
Acta 1381 (1998) 319^330.
[12] C.V. Hall, C. Yanofsky, J. Bacteriol. 148 (1981) 941^949.
[13] J. Scha«ferjohann, J.-G. Yoo, B. Kusian, B Bowien, J. Bac-
teriol. 175 (1993) 7329^7340.
[14] J.L. Gibson, F.R. Tabita, J. Bacteriol. 179 (1997) 663^669.
[15] J.L. Gibson, F.R. Tabita, Arch. Microbiol. 166 (1996) 141^
150.
[16] B. Kusian, B. Bowien, FEMS Microbiol. Rev. 21 (1997)
135^155.
[17] J.H. Miller, A Short Course in Bacterial Genetics, Cold
Spring Harbor Laboratory Press, Cold Spring Harbor,
NY, 1992.
[18] D.J. Clark, O. MaalÖe, J. Mol. Biol. 23 (1967) 99^112.
[19] B. Berlin, Ph.D. Thesis, 1993.
[20] J.R. Scott, Microbiol. Rev. 48 (1984) 1^23.
[21] M.M. Bradford, Anal. Biochem. 72 (1976) 248^254.
[22] A. Goldfarb, Proc. Natl. Acad. Sci. USA 78 (1981) 3454^
3458.
[23] F.J. Calzone, R.J. Britten, E.H. Davidson, in: S.L. Berger,
A.R. Kimmel (Eds.), Methods in Enzymology: Guide to
Molecular Cloning Techniques, Academic Press, San Diego,
1987, pp. 611^632.
[24] R.D. Fleischmann, M.D. Adams, O. White, R.A. Clayton,
E.F. Kirkness, A.R. Kerlavage, C.J. Bult, J.F. Tomb, B.A.
Dougherty, J.M. Merrick et al., Science 269 (1995) 496^512.
[25] W. Martin, C. Schnarrenberger, Curr. Genet. 32 (1997) 1^
18.
[26] L.N. Ornston, M.K. Ornston, J. Bacteriol. 98 (1969) 1098^
1108.
[27] D.P. Clark, J.E.J. Cronan, in: F.C. Neidhardt (Ed.), Escher-
ichia coli and Salmonella: Cellular and Molecular Biology,
American Society for Microbiology, Washington, DC, 1996,
pp. 343^357.
384 A. Lyngstadaas et al. / Biochimica et Biophysica Acta 1472 (1999) 376^384
[28] F. Kunst, N. Ogasawara, I. Moszer, A.M. Albertini, G.
Alloni, V. Azevedo, M.G. Bertero, P. Bessieres, A. Bolotin,
S. Borchert, R. Borriss, L. Boursier, A. Brans, M. Braun,
S.C. Brignell, S. Bron, S. Brouillet, C.V. Bruschi, B. Cald-
well, V. Capuano, N.M. Carter, S.K. Choi, J.J. Codani, I.F.
Connerton, A. Danchin et al., Nature 390 (1997) 249^256.
[29] S. Reverchon, D. Expert, J. Robert-Baudouy, W. Nasser,
J. Bacteriol. 179 (1997) 3500^3508.
[30] R. Rajagopalan, W. Altekar, Eur. J. Biochem. 221 (1994)
863^869.
[31] R. Snyder, W. Champness, Molecular Genetics of Bacteria,
American Society for Microbiology, Washington, DC, 1997.
[32] A.J. Pittard, in: F.C. Neidhardt (Ed.), Escherichia coli and
Salmonella : Cellular and Molecular Biology, American Soci-
ety of Microbiology, Washington, DC, 1996, pp. 458^484.
[33] D. Vinella, B. Gagny, D. Joseleau-Petit, R. D'Ari, M. Cash-
el, J. Bacteriol. 178 (1996) 3818^3828.
[34] E. Boye, A. LÖbner-Olesen, Cell 62 (1990) 981^989.
[35] P. Modrich, Annu. Rev. Microbiol. 25 (1991) 229^253.
[36] J. Plumbridge, D. Soll, Biochimie 69 (1987) 539^541.
[37] M.W. van der Woude, B.A. Braaten, D.A. Low, Mol. Mi-
crobiol. 6 (1992) 2429^2435.
[38] D. Roberts, B.C. Hoopes, W.R. McClure, N. Kleckner, Cell
43 (1985) 117^130.
BBAGEN 24895 1-10-99
[39] Y. Sakakibara, Mol. Microbiol. 24 (1997) 793^801.
[40] L.J. Rasmussen, M.G. Marinus, A. LÖbner-Olesen, Mol.
Microbiol. 12 (1994) 631^638.
[41] L.J. Rasmussen, A. LÖbner-Olesen, M.G. Marinus, Gene
157 (1995) 213^215.
[42] R.E.J. Wolf, D.M. Prather, F.M. Shea, J. Bacteriol. 139
(1979) 1093^1096.
[43] W. Ross, K.K. Gosink, J. Salomon, K. Igarashi, C. Zou, A.
Ishihama, K. Severinov, R.L. Gourse, Science 262 (1993)
1407^1413.
[44] R.L. Gourse, T. Gaal, M.S. Bartlett, J.A. Appleman, W.
Ross, Annu. Rev. Microbiol. 50 (1996) 645^677.
[45] C.V. Hall, C. Yanofsky, J. Bacteriol. 151 (1982) 918^923.
[46] H.-C. Tsui, G. Zhao, G. Feng, H.-C. Leung, M.E. Winkler,
Mol. Microbiol. 11 (1994) 189^202.
[47] L.J. Rasmussen, P.L. MÖller, T. Atlung, J. Bacteriol. 173
(1991) 6390^6397.
[48] B.J. Bachmann, in: F.C. Neidhardt (Ed.), Escherichia coli
and Salmonella : Cellular and Molecular Biology, American
Society for Microbiology, Washington, DC, 1996, pp. 2460^
2488.
[49] B. Parker, M.G. Marinus, Gene 73 (1988) 531^535.
[50] T. Linn, G. Ralling, Plasmid 14 (1985) 134^142.