![Palmitate labeling of NlpE. (A) Cellular fractions from cultures transformed with the parental plasmid pBAD18. (B) Cellular fractions from cultures transformed with a nlpE-producing plasmid, pND18. Whole-cell (W.C.), soluble cytoplasmic and periplasmic contents (Sol.), total-membrane (T.M.), inner membrane (I.M.), and outer membrane (O.M.) fractions were created from [ 3 H]palmitate-labeled cultures and subjected to SDS-PAGE (see Materials and Methods).](https://www.researchgate.net/profile/Christine-Cosma/publication/15562957/figure/fig1/AS:341545699037184@1458442248116/Palmitate-labeling-of-NlpE-A-Cellular-fractions-from-cultures-transformed-with-the_Q320.jpg)
Content uploaded by Christine L Cosma
Author content
All content in this area was uploaded by Christine L Cosma on Jan 27, 2014
Content may be subject to copyright.
JOURNAL OF BACTERIOLOGY, Aug. 1995, p. 4216–4223 Vol. 177, No. 15
0021-9193/95/$04.0010
Copyright 1995, American Society for Microbiology
Overproduction of NlpE, a New Outer Membrane Lipoprotein,
Suppresses the Toxicity of Periplasmic LacZ by Activation
of the Cpx Signal Transduction Pathway
WILLIAM B. SNYDER, LAURA J. B. DAVIS,† PAUL N. DANESE,
CHRISTINE L. COSMA, AND THOMAS J. SILHAVY*
Department of Molecular Biology, Princeton University, Princeton, New Jersey 08544
Received 16 February 1995/Accepted 14 May 1995
The LamB-LacZ-PhoA tripartite fusion protein is secreted to the periplasm, where it exerts a toxicity of
unknown origin during high-level synthesis in the presence of the inducer maltose, a phenotype referred to as
maltose sensitivity. We selected multicopy suppressors of this toxicity that allow growth of the tripartite fusion
strains in the presence of maltose. Mapping and subclone analysis of the suppressor locus identified a
previously uncharacterized chromosomal region at 4.7 min that is responsible for suppression. DNA sequence
analysis revealed a new gene with the potential to code for a protein of 236 amino acids with a predicted
molecular mass of 25,829 Da. The gene product contains an amino-terminal signal sequence to direct the
protein for secretion and a consensus lipoprotein modification sequence. As predicted from the sequence, the
suppressor protein is labeled with [
3
H]palmitate and is localized to the outer membrane. Accordingly, the gene
has been named nlpE (for new lipoprotein E). Increased expression of NlpE suppresses the maltose sensitivity
of tripartite fusion strains and also the extracytoplasmic toxicities conferred by a mutant outer membrane
protein, LamBA23D. Suppression occurs by activation of the Cpx two-component signal transduction pathway.
This pathway controls the expression of the periplasmic protease DegP and other factors that can combat
certain types of extracytoplasmic stress.
The analysis of protein secretion in the gram-negative bac-
terium Escherichia coli has been aided by gene fusion technol-
ogy (2, 24). Hybrid proteins containing amino-terminal do-
mains derived from secreted periplasmic or outer membrane
proteins fused to an active LacZ (b-galactosidase) domain
have proven useful for studying secretion. Fusion of MalE
(periplasmic maltose-binding protein) or LamB (outer mem-
brane maltoporin or lreceptor) to the normally cytoplasmic
protein LacZ confers a toxicity upon host strains during con-
ditions of high-level synthesis. Strains harboring these fusions
lyse during growth in the presence of the inducer maltose, a
phenotype referred to as maltose sensitivity (1, 26). Intragenic
mutations that relieve the toxicity of these proteins alter the
amino-terminal signal sequence and prevent entry into the
secretory pathway (8). By using this approach, the essential
features of the signal sequence were defined. These hybrid
proteins also show abnormally low levels of LacZ activity un-
der noninducing conditions. The low activity presumably re-
sults from secretion of the hybrid protein to the periplasm,
where it is inactivated by intermolecular disulfide bond forma-
tion (27). Hence, signal sequence mutations increase LacZ
activity by preventing secretion and subsequent inactivation.
Additionally, extragenic mutations that impair secretion in-
crease retention of the hybrid protein in the cytoplasm, result-
ing in increased LacZ activity. This approach identified several
of the sec genes which encode components of the secretory
machinery (2, 24).
New variants of the standard lamB-lacZ hybrid that bestow
novel phenotypes were constructed (27). The maltose sensitiv-
ity conferred by the original LamB-LacZ hybrid protein is
exerted in the cytoplasm and results from an inhibition of the
secretion machinery during high-level synthesis of the hybrid
protein. This consequence of LamB-LacZ overproduction is
called hybrid protein jamming. Truncation of the LacZ domain
by fusion of PhoA or by introduction of the late nonsense
mutation X90 has created a new class of model proteins,
LamB-LacZ-PhoA and LamB-LacZX90, that do not inhibit
the secretion machinery at growth temperatures of 348Cor
higher. Even though these proteins do not jam, they impart a
maltose sensitivity that is indistinguishable from that of LamB-
LacZ. This novel maltose sensitivity results from secretion of
the X90 and tripartite fusion proteins to the periplasm, where
they form disulfide-bonded aggregates.
Another novel envelope associated toxicity is conferred by a
mutant outer membrane protein, LamBA23D, that contains a
defect in the signal sequence cleavage site (4). The precursor
form of LamBA23D is poorly processed by leader peptidase
(LepA) and accumulates in the bacterial envelope. The accu-
mulation of unprocessed LamBA23D creates a permeability
defect as evidenced by increased sensitivity to detergent (4)
and the antibiotic amikacin (5a). The cause of this toxicity may
be related to targeting of the mutant precursor to the outer
membrane, but it is not known how this leads to manifestation
of the observed defects.
To learn more about the nature of the extracytoplasmic
toxicities of LamBA23D, LamB-LacZX90, and LamB-LacZ-
PhoA, suppressors of their toxicities were selected. A class of
suppressors that combats the toxic effects of all these proteins
was discovered. These suppressor mutations encode activated
alleles of the previously characterized gene cpxA and conse-
quently have been referred to as cpxA* (18). CpxA and CpxR
form the sensor and response regulator components of the
well-characterized family of bacterial two-component regula-
tory systems (7, 32). These systems utilize an inner membrane
protein to sense an environmental parameter and communi-
* Corresponding author. Phone: (609) 258-5899. Fax: (609) 258-
6175.
† Present address: University of Texas, Southwestern Medical
School, Dallas, TX 75235-9096.
4216
cate the information via a kinase/phosphatase activity to a
response regulator (29). The response regulator is often a
DNA-binding protein that can elicit an alteration in transcrip-
tion. cpxA* alleles function in part by increasing the expression
of degP (6), the gene specifying a major periplasmic protease
(17, 30). Increased synthesis of DegP and another unknown
factor(s) renders cpxA* strains resistant to high-level synthesis
of the toxic envelope proteins (5a). Accordingly, CpxA is
thought to control the expression of genes that can combat
extracytoplasmic stress.
Here we describe the isolation of a previously uncharacter-
ized gene which can confer maltose resistance to lamB-lacZ-
phoA strains when it is expressed from a high-copy-number
plasmid. Reminiscent of cpxA* suppressors, high-level expres-
sion of this gene can also combat the toxicities exhibited by
lamBA23D strains. These results provide new information
about the functioning of the Cpx signal transduction pathway.
MATERIALS AND METHODS
Bacterial strains and plasmids. The E. coli K-12 strains used in this study are
all derivatives of strain MC4100 (5). The parent of tripartite fusion strains is
WBS1106 {MC4100 DphoA532 F(lamB-lacZ-phoA) Hyb1-1 [lp1(209)]} (27).
The parent of lamBA23D strains is JHC285 (4). The cpxR::spc allele (6) and the
degP::Tn10 allele (laboratory collection) were transduced into different back-
grounds by P1. Strain DH5awas used as a host for plasmids during cloning
procedures (23). The drpA plasmid, pKLF2, is described elsewhere (33). The
plasmid library used in this study was obtained from Susan Gottesman and
contains Sau3A partially digested chromosomal DNA from a MC4100Dlon strain
cloned at the BamHI site of pBR322. Plasmids pLD130 and pLD111 are two of
the identical maltose-resistant suppressor plasmids isolated from the library. Six
other suppressor plasmids were analyzed, but all of them appeared identical to
pLD130 and pLD111 following restriction analysis. pND18, which contains nlpE
cloned under the control of the arabinose promoter in the vector pBAD18, is
described elsewhere (6).
Media and chemicals. All growth media used have been described previously
(25) with the exception of glycerol minimal medium, which contained 0.4%
glycerol and 0.5% Luria broth. We purchased [
3
H]palmitate (1 mCi/0.2 ml, 55
Ci/mM) from NEN Research Products, Du Pont Co., Boston, Mass., and Kodak
XAR film from Eastman Kodak, Rochester, N.Y. 5-Bromo-4-chloro-3-in-
dolylphosphate (X-Phos) was purchased from Calbiochem, LaJolla, Calif. The E.
coli gene-mapping membrane was purchased from TaKaRa Biomedical, Shiga,
Japan. Difco Laboratories, Detroit, Mich., and BBL, Cockeysville, Md., supplied
antibiotic disks for sensitivity tests. Reagents for enhanced chemiluminescence
detection, random-prime labeling, and DNA hybridization were obtained from
Amersham, Arlington Heights, Ill. All DNA restriction enzymes, b-agarase I, T4
DNA ligase, and Klenow polymerase were purchased from New England Bio-
labs, Beverly, Mass. Reagents for sequence analysis were obtained from United
States Biochemical, Cleveland, Ohio.
DNA techniques. Competent cells were prepared for plasmid transformation
by standard techniques. Sequence analysis of pLD404 was performed with the
EcoRI-StyI DNA fragment from pLD404. Purification and sequencing of this
fragment were performed as described previously for PCR products (20, 22).
Oligonucleotide primers were obtained from the Princeton University Syn/Seq
Facility for sequencing and PCR amplification of DNA. For mapping analysis, a
randomly primed DNA probe of pLD404 was prepared, hybridized, washed, and
detected on the gene-mapping membrane by enhanced chemiluminescence as
described by the supplier of these reagents. PCR amplification of DNA from a
single bacterial colony has been described previously (22). Subclones of pLD130
were created by digestion with the restriction enzyme whose sites defined the
boundaries of the region to be deleted. The plasmid was then ligated by standard
techniques to create the deletions.
Construction of the nlpE null mutation. 59and 39regions of nlpE were
amplified by PCR with primers containing restriction sites for cloning into a
vector with a temperature-sensitive replication origin (pMAK705) (12). The
primers for amplification of the 59region were NLPEF (59GGAATTCCGAC
GACCCACGC39) and DNLP1 (59CGGATCCAGCTTGTCAGCGG39). The
primers for amplification of the 39region were DRPABAK (59CCGCGTCTT
CAGCACTTCC39) and DNLP2 (59CGGATCCCGATACGGCAGGG39).
pMAK705 was partially cut with EcoRI and then cut to completion with BamHI.
The large plasmid from this digest was gel purified as described previously (20)
and ethanol precipitated by standard techniques, as were all DNA fragments
prior to ligation. DNA amplified from the 59region of nlpE was cut with EcoRI
and BamHI and purified. These fragments were then ligated under standard
conditions and transformed into DH5aat 308C. This construct, pBS6, and DNA
amplified from the 39region were cut with BamHI and HindIII, purified, ligated,
and transformed. This construct, pBS7, and the spc interposon-containing plas-
mid, pHP45V(9), were cut with BamHI, and the appropriate fragments were
purified, ligated, and transformed. This final construct, pBS8, contains the polar
spc interposon flanked by nlpE DNA. This plasmid was transformed into
MC4100 at 428C by selecting resistance to spectinomycin to ensure cointegrate
formation. The transformants were purified three times at 428C and screened for
cold resistance and ampicillin sensitivity to confirm resolution of the cointegrate.
The chromosomal deletion/insertion at nlpE was confirmed by PCR amplifica-
tion and restriction enzyme analysis. P1-mediated generalized transduction was
used to move this allele into different genetic backgrounds.
Disk sensitivity assays. Strains for maltose sensitivity tests were grown to
saturation in glycerol minimal medium containing ampicillin (50 mg/ml), pel-
leted, and resuspended in 1/2 volume of M63. Maltose sensitivity was measured
against 10 ml of maltose solution (concentrations are given in Tables 1 and 2) on
a filter paper disk atop a lawn of the test strain (100 ml) in F-top agar (3 ml) on
glycerol minimal medium plates. The antibiotic and detergent sensitivities of the
nlpE::spc and isogenic nlpE
1
strains were compared on Luria broth agar with
lawns of cells (100 ml) in Luria broth top agar (3 ml). Strains for this test were
grown to saturation in Luria broth and added directly to the top agar. The
following drug-containing disks (Difco and BBL) and detergents (10 ml per disk)
were used for this analysis: chloramphenicol, kanamycin, gentamicin, rifampin,
sulfathiazole, moxalactam, penicillin G, nafcillin, oxytetracycline, bacitracin, no-
vobiocin, neomycin, cephalothin, clindamycin, chlorotetracycline, tetracycline,
trimethoprin, amikacin, polymyxin B, cloxacillin, naladixic acid, 10% sodium
dodecyl sulfate (SDS), and 20% deoxycholate. All assays were performed at
378C. Results are expressed as the diameter of growth inhibition after subtrac-
tion of the diameter of the filter paper disk in millimeters. Standard concentra-
tions of antibiotics were included in the top agar to maintain plasmids when
appropriate.
Palmitate labeling. Overnight cultures were diluted 1:50 into 50 ml of glycerol
minimal medium containing ampicillin and grown at 378CtoanA
600
of approx-
imately 0.2 to 0.3. Arabinose was then added to a final concentration of 0.4%.
The cultures were incubated for 15 min at 378C, 250 mCi of [
3
H]palmitate (50 ml)
was added, and the cultures were grown for a further3hat378C. The cultures
were then collected on ice.
Isolation and fractionation of total membranes. A 50-ml portion of either
palmitate-labeled or unlabeled cells was pelleted, resuspended in 25 ml of 50 mM
Tris (pH 7.5) (buffer), repelleted, and suspended in 2 ml of buffer with the
following additives: 2 ml of RNase I (2 mg/ml), 2 ml of DNase I (1 mg/ml), 2 ml
of leupeptin (5 mg/ml), 4 ml of aprotinin (10 mg/ml), 1 ml of pepstatin A (1 M),
and 20 ml of phenylmethylsulfonyl fluoride (100 mM in ethanol) (all from
Sigma). The cells were lysed by two passages through a French pressure cell press
(diameter, 3/8-in. [0.95 cm]) at 15,000 lb/in
2
. Samples were centrifuged at 5,000
rpm for 15 min in a Beckman SS34 rotor to pellet unlysed cells and debris. The
supernatant was centrifuged in a TLA100.2 rotor in a Beckman Optima TL
ultracentrifuge at 100,000 rpm for 20 min. The supernatant was saved as a soluble
fraction; the pellet made up the total membrane fraction. The membranes were
resuspended in 100 ml of buffer and loaded onto a sucrose step gradient con-
sisting of 0.3 ml of 70% and 0.75 ml of 53% sucrose–50 mM Tris (pH 7.5). The
gradients were centrifuged at 100,000 rpm for 65 min in the TLA100.2 rotor with
the acceleration and deceleration set to 5. We observed bands of inner (upper)
and outer (lower) membrane in the gradient following centrifugation and col-
lected them as 100-ml fractions. All manipulations were performed on ice and at
48C whenever possible.
SDS-polyacrylamide gel electrophoresis (PAGE) and autoradiography. Previ-
ously described methods were used (28). The volume of loading buffer added per
milliliter of pelleted cells was calculated by the following formula: volume (mil-
liliters) 5A
600
/5. Samples prepared following isolation of membrane fractions
were resuspended in loading buffer to approximate the dilution obtained by the
above formula. The inner and outer membrane samples obtained from the
fractionation procedure were resuspended in an equal volume of loading buffer
that contained no glycerol.
Nucleotide sequence accession number. The nucleotide and deduced amino
acid sequences of nlpE have been deposited in GenBank under accession num-
ber U18345.
RESULTS
Rationale. High-level synthesis of LamB-LacZ-PhoA during
growth in the presence of the inducer maltose exerts a periplas-
mic toxicity that results in cell lysis. This protein forms high-
molecular-weight aggregates in the periplasm of dying cells,
but the precise cause of toxicity remains a mystery. Perhaps
periplasmic LacZ titrates some essential periplasmic func-
tion(s). Expression of the gene encoding this titrated function
from a multicopy plasmid might overcome lethality by provid-
ing an excess of the inactivated function. Alternatively, pro-
teins that can combat the toxic effects of this periplasmic pro-
tein could be identified in the same manner. Therefore,
multicopy suppression of the maltose-sensitive lamB-lacZ-
VOL. 177, 1995 NlpE OVERPRODUCTION SUPPRESSES PERIPLASMIC LacZ 4217
phoA strain should provide clues about the nature of this
toxicity or how cells cope with periplasmic stress.
Selection and identification of a suppressor plasmid. We
performed multicopy suppression with a library of cloned chro-
mosomal DNA. This library (Sau3A partially digested chro-
mosomal DNA cloned into the BamHI site of pBR322) was
transformed into the maltose-sensitive lamB-lacZ-phoA strain,
WBS1106. Following phenotypic expression, the cells were
plated directly onto maltose minimal agar containing ampicil-
lin and X-Phos, a chromogenic indicator of alkaline phos-
phatase (PhoA) activity. After 2 days of incubation at 378C,
lawns of transformants grew on this agar because of the ability
of the strain to initially grow on maltose medium. Replica
printing of these lawns onto maltose minimal agar containing
ampicillin and X-Phos provided additional selection for mal-
tose-resistant mutants. Mutations that prevent synthesis or se-
cretion of LamB-LacZ-PhoA appeared white on the indicator
medium, thus providing an easy way to avoid this class of
maltose-resistant suppressors. Following purification of 91
maltose-resistant and PhoA
1
isolates, plasmid DNA was pre-
pared and transformed into WBS1106 to confirm the plasmid
linkage of the suppressor. A total of 76 plasmid-borne suppres-
sors were identified in this manner. Approximately 1 in 2,400
transformants contained a suppressor plasmid.
Eight plasmids that conferred maltose resistance were sub-
jected to restriction enzyme analysis to determine if they had
common restriction fragments. All eight plasmids examined in
this manner contained the same restriction fragments (data not
shown). Further restriction analysis of one of these plasmids
(pLD130) was performed, and a map showing the approximate
location of relevant restriction sites is shown in Fig. 1.
Mapping. We determined the physical map location of the
cloned DNA from a suppressor plasmid by DNA hybridization
techniques. A gene-mapping membrane containing DNA from
the overlapping, ordered clones of the Kohara phage library is
available commercially. Randomly primed labeled DNA was
created from a suppressor plasmid and hybridized to the gene-
mapping membrane (see Materials and Methods). Plasmid-
derived DNA hybridized to Kohara phages 122 and 123 (14),
localizing the suppressor gene at 4.7 to 4.9 min on the chro-
mosome (21). The similarity between restriction sites of the
plasmid insert and those reported for this genomic region (21)
confirmed the map location.
These mapping data suggested that our suppressor plasmid,
pLD130, probably contained the previously characterized drpA
gene (33). A drpA temperature-sensitive mutant and the orig-
inal drpA plasmid were obtained to determine if this gene was
responsible for suppression. Unlike the suppressor plasmid,
the original drpA plasmid, pKLF2, does not contain any DNA
extending to the left of the HindIII (H) site (Fig. 2A). The
temperature sensitivity of the drpA strain was complemented
by pLD130 as well as pKLF2. However, pKLF2 did not confer
maltose resistance; consequently, this provides evidence that
the region to the left of the HindIII site is responsible for
suppression by pLD130 (Fig. 2A).
To demonstrate directly that the region downstream of
drpA, to the left of the HindIII site, is responsible for suppres-
sion, we created subclones of pLD130. Subclones were created
FIG. 1. Restriction map of pLD130. Important restriction sites of the
genomic insert (shaded band) and plasmid backbone are shown. Positions of sites
are approximate.
FIG. 2. Restriction map and suppressor activity of plasmid clones. (A) The subclones pLD401, pLD402, and pLD404 were created by removal of the DNA
represented by the gaps from pLD130 and religating. pKLF2 is the original drpA clone (33). Abbreviations of restriction enzyme sites: B, BamHI; E, EcoRI; H, HindIII;
S, StyI. The effect on maltose sensitivity was determined in WBS1106. Mal
R
, no growth inhibition was observed in the disk assay (see Materials and Methods); Mal
S
,
a wild-type zone of growth inhibition was observed. (B) For clarity, the approximate positions of genes in this chromosomal region are shown (see the text).
4218 SNYDER ET AL. J. BACTERIOL.
by deleting parts of the genomic insert (see Materials and
Methods), and the ability of each to confer maltose resistance
was assayed (Fig. 2A). One subclone, pLD404, containing ap-
proximately 1,800 bp of DNA downstream of drpA, conferred
the suppressor phenotype and was subjected to DNA sequence
analysis. This region contains the complete coding sequence
for a new gene that we are calling nlpE and two other partial
open reading frames (ORFs) described below (Fig. 2B).
Sequence analysis of the suppressor region. Both strands of
DNA from the insert region of pLD404 were sequenced by the
dideoxy method (Fig. 3A). The sequence of the insert contains
three significant ORFs that could potentially encode a gene
product responsible for suppression (Fig. 3). First, an ORF
downstream of the drpA gene spans the HindIII and StyI re-
striction sites (Fig. 3, 9yaeF [ORF292]). The 59region of this
ORF is found on the drpA sequence (33) (Fig. 2); therefore,
the complete gene is not present in our suppressor plasmid. It
is therefore likely that one of the other ORFs mediates sup-
pression. A complete ORF of 236 amino acids (nlpE, see
below) is located downstream of a partial ORF of 124 amino
acids at the other end of the genomic insert. In the accompa-
nying paper by Gupta et al., the partial ORF is reported to be
composed of 140 amino acids and as such will be called
ORF140 (11). The proximity of the stop codon of ORF140 and
the translational start site of nlpE, as well as the lack of any
obvious promoter in this region, suggest that the two genes are
normally found in an operon. The 59end of this operon is
absent in the suppressor plasmid, but ORF140 is situated such
that the tetracycline resistance protein (TetA) from pBR322 is
fused to it in frame (Fig. 2 and 3B). This tetA hybrid gene is
transcribed constitutively from the tetA promoter (3, 31), thus
providing transcription of nlpE.
Since there was only one complete ORF present on our
plasmid, nlpE, we focused our attention on this gene. Close
examination of the translated protein sequence revealed sev-
eral interesting features (10). First, the protein appeared to
contain an amino-terminal signal sequence to direct it for
secretion from the cytoplasm through the general secretory
pathway. Second, the protein contains a lipoprotein modifica-
tion sequence of LMGC, at the carboxy-terminal end of the
presumed signal sequence. This sequence defines the cleavage
site for signal peptidase II (lsp); cleavage occurs between G
and C. This cysteine is modified with a diacylglyceryl moiety
prior to the cleavage of lipoprotein precursors by signal pep-
tidase II, thus resulting in a mature protein with diacylglyceryl-
cysteine at the amino terminus (13). Such cleavage products
undergo further posttranslational modification at the amino
terminus, i.e., N-acylation (13). If these sequences in NlpE are
functional, the protein will be secreted to the envelope with
incorporated palmitate. The toxicity of LamB-LacZ-PhoA is
exerted in the envelope, the presumed location of NlpE. Ac-
cordingly, nlpE became the primary candidate for mediating
suppression by pLD404.
Identification of the suppressor protein. To facilitate iden-
tification of the suppressor gene, we cloned nlpE into the
arabinose-inducible expression vector pBAD18. The construc-
tion of this plasmid (pND18), which encodes nlpE, is described
elsewhere (6). This plasmid confers maltose resistance to
lamB-lacZ-phoA-containing strains, thus proving that in-
creased expression of nlpE mediates suppression (Table 1).
The amount of arabinose used to induce expression is critical
to the ability of this plasmid to suppress. Maltose sensitivity is
not suppressed during conditions of high-level expression
(0.2% arabinose) or low-level synthesis (0% arabinose) (Table
1). The level of NlpE that is produced from our original sup-
pressor plasmid (pLD130) was fortuitously appropriate for
suppression.
To prove that the lipoprotein modification sequence of nlpE
is functional, we labeled cells with [
3
H]palmitate to see if
pND18 overexpressed a palmitate-modified protein. Cells con-
taining either pND18 or the control plasmid pBAD18 were
labeled with [
3
H]palmitate to specifically label cellular lipopro-
teins as described in Materials and Methods. Whole-cell, sol-
uble, total-membrane, inner membrane, and outer membrane
fractions from each strain reveal the subcellular distribution of
the overexpressed protein (see Materials and Methods). Fig-
ure 4 shows that the pND18-containing cells overexpress a
protein that incorporates [
3
H]palmitate. This protein has an
apparent molecular mass of 25,000 Da, which is in close agree-
ment with the predicted molecular mass of 23,722 Da for the
mature protein. Both the inner and outer membrane fractions
contain this protein when overproduced, as well as several
apparent degradation products. In contrast, a rare lipoprotein
with the same apparent molecular weight as the overexpressed
protein is found only in the outer membrane fraction of the
control vector cells. These data suggest that the gene respon-
sible for suppression encodes a rare outer membrane lipopro-
tein. We propose that this gene be named nlpE, for new li-
poprotein E.
A mechanism of suppression. We have previously reported
that NlpE overproduction provides a strong signal to induce
the expression of degP and that induction requires functional
cpxA and cpxR (6). Suppression of the maltose sensitivity of
lamB-lacZ-phoA strains by NlpE overproduction absolutely re-
quires cpxR (Table 2). CpxA and CpxR form the sensor and
response regulator components of the well-characterized bac-
terial two-component regulatory systems (see Introduction).
Therefore, NlpE multicopy suppression functions by activation
of the Cpx two-component regulatory system.
Previous work from our laboratory has identified cpxA* al-
leles as genetic suppressors of LamB-LacZ-PhoA- and LamBA
23D-mediated toxicities (see Introduction). These cpxA* alle-
les completely abolish the extracytoplasmic toxicities conferred
by these model proteins. With the tripartite fusion, these sup-
pressors function by increasing the expression of the periplas-
mic protease DegP, which enhances the degradation of the
toxic protein (5a). We thought it likely that overproduction of
NlpE would function by a similar mechanism, since suppres-
sion is dependent on cpxR. Consistent with this hypothesis,
NlpE production from a suppressor plasmid enhances the deg-
radation of LamB-LacZ-PhoA (data not shown). As with
cpxA* suppression, enhanced degradation requires degP; how-
ever, considerable suppression of the maltose sensitivity of
lamB-lacZ-phoA strains occurs in a degP null background (Ta-
ble 2). This suggests the presence of other factors, besides
DegP, that are regulated by the Cpx pathway to suppress mal-
tose sensitivity.
We have also found that NlpE overproduction significantly
decreases the SDS- and amikacin-sensitive phenotypes of
lamBA23D strains. Moreover, this effect is cpxR dependent.
NlpE therefore mediates suppression through the Cpx pathway
by regulating the expression of factors that can combat these
extracytoplasmic toxicities.
Induction of DegP does not always confer suppression.
Overproduction of outer membrane proteins, as well as other
insults to the bacterial envelope, is known to increase degP
synthesis (6, 19). This induction of degP, however, does not
function through the Cpx pathway (6). We tested the ability of
overexpression of the outer membrane protein OmpF, which
strongly induces degP, to suppress the maltose sensitivity of
lamB-lacZ-phoA strains. We observed no differences in the
VOL. 177, 1995 NlpE OVERPRODUCTION SUPPRESSES PERIPLASMIC LacZ 4219
FIG. 3. DNA sequence and ORFs from the chromosomal region responsible
for suppression. (A) The DNA sequence, ORFs, and other relevant features of
the uncharacterized region from pLD404 (GenBank accession number, U18345).
The coding strand of nlpE is shown. The complementary strand codes for yaeF
(ORF292). SPIM marks the start of the serine protease inhibitor motif (10), and
S.D. marks the position of the Shine-Dalgarno ribosome-binding site. The un-
derlined amino acids mark the lipoprotein modification sequence. (B) Illustra-
tion of the orientation of the ORFs with respect to plasmid-encoded genes. All
positions are approximate.
4220 SNYDER ET AL. J. BACTERIOL.
diameters of growth inhibition by maltose for a lamB-lacZ-
phoA fusion strain transformed with the OmpF plasmid or a
control vector. These results demonstrate that the induction of
degP by this mechanism is insufficient to combat the toxicity of
periplasmic LacZ.
Characterization of an nlpE null mutant. Because NlpE-
mediated suppression works through the Cpx signal transduc-
tion pathway, NlpE may normally function in this pathway. To
test the normal, wild-type function of NlpE, we created a null
mutation in the gene. A disrupted copy of nlpE was con-
structed on a plasmid that could be integrated into the chro-
mosome at nlpE. Resolution of this cointegrate leaves the
disrupted copy of nlpE in the chromosome (see Materials and
Methods). The resulting nlpE::spc allele has the potential to
code for the first 98 amino acids of the NlpE precursor protein.
However, the deletion/substitution caused by the spectinomy-
cin interposon (9) removes a significant portion of the coding
sequence. This disruption has been confirmed by PCR analysis
and presumably creates a null allele of nlpE.
Characterization of the nlpE::spc mutant strain has not re-
vealed any remarkable phenotypes. The nlpE::spc mutation
does not confer noticeable growth defects during growth on all
media tested at any temperature (23 to 448C). The maltose
sensitivity of lamB-lacZ-phoA strains is not affected by the nlpE
null mutation. No differences were observed in the sensitivities
of the nlpE::spc strain to 21 different antibiotics, SDS, and
deoxycholate (see Materials and Methods), thus providing ev-
idence that the mutation does not alter the permeability prop-
erties of cells. We prepared whole-cell, soluble, total-mem-
brane, inner membrane, and outer membrane fractions from
the isogenic nlpE::spc and nlpE
1
strains and found no differ-
ences following SDS-PAGE of these samples (data not shown).
Also, we observed no effect of the nlpE::spc insertion on the
induction of degP in cpxA* mutant backgrounds or with other
treatments that are known to induce degP, such as outer mem-
brane protein overproduction (data not shown). This analysis
does not rule out the possibility that NlpE functions to stimu-
late the Cpx pathway under certain conditions. However, since
no conditions other than NlpE overproduction are known to
stimulate the Cpx pathway, tests of epistasis are impossible at
this time. The accompanying paper by Gupta et al. suggests a
role for this newly identified lipoprotein in copper transport
and homeostasis (11).
DISCUSSION
The LamB-LacZ-PhoA fusion protein is secreted efficiently
to the periplasm, where it exerts a pronounced cellular toxicity
(27). We suspect that in this location, LacZ titrates some es-
sential periplasmic function(s). To identify genes whose prod-
ucts can combat this toxicity when expressed at elevated levels,
we used the technique of multicopy suppression. Using this
method, we have identified a new gene, nlpE, that maps to 4.7
min and encodes a rare 24-kDa outer membrane lipoprotein.
We have shown that NlpE overproduction causes suppression
by activating the Cpx two-component signal transduction path-
way. This regulatory system contains the membrane receptor
kinase CpxA and the response regulator CpxR (see Introduc-
tion). degP is one of the known downstream targets in the Cpx
pathway (6), and tests of epistasis confirm that all of these gene
products are important for suppression. Increased levels of this
periplasmic protease enhance the degradation of the LamB-
LacZ-PhoA fusion protein, thus preventing its accumulation to
toxic levels. Mutations that activate the Cpx pathway (cpxA*)
suppress the toxicity associated with periplasmic LacZ by a
similar mechanism (5a).
A mutation such as lamBA23D, which hinders the removal
of the signal sequence from an outer membrane precursor
protein, confers growth defects as well (4). In this case, we
FIG. 4. Palmitate labeling of NlpE. (A) Cellular fractions from cultures
transformed with the parental plasmid pBAD18. (B) Cellular fractions from
cultures transformed with a nlpE-producing plasmid, pND18. Whole-cell (W.C.),
soluble cytoplasmic and periplasmic contents (Sol.), total-membrane (T.M.),
inner membrane (I.M.), and outer membrane (O.M.) fractions were created
from [
3
H]palmitate-labeled cultures and subjected to SDS-PAGE (see Materials
and Methods).
TABLE 1. NlpE confers maltose resistance to
lamB-lacZ-phoA strains
a
Plasmid and arabinose
concn (%) in culture
b
Diam of growth inhibition (mm) in
following arabinose concn (%)
in top agar
c
:
0 0.05 0.2
pND18
0 9 10 10
0.05 8 0 8
d
0.2 9 14 11
pBAD18
0ND
e
ND ND
0.05 11 10 8
0.2 11 9 9
a
Sensitivity was quantitated as described in Materials and Methods for 2%
maltose. Results from a typical experiment are shown.
b
Overnight cultures of the test strains were grown in the given concentration
of arabinose.
c
Test strains were plated in F-top agar containing the given concentration of
arabinose.
d
A hazy zone of growth inhibition was observed; all other zones were clear.
e
ND, not determined.
TABLE 2. NlpE-mediated suppression is dependent
on cpxR and degP
a
Relevant genotype Diam of growth inhibition (mm) in:
2% maltose 10% maltose
pLD404 (nlpE
1
)
lamB-lacZ-phoA 00
lamB-lacZ-phoA cpxR::spc 14 21
lamB-lacZ-phoA degP::Tn10 012
pBR322 (control)
lamB-lacZ-phoA 916
lamB-lacZ-phoA cpxR::spc 11 19
lamB-lacZ-phoA degP::Tn10 10 17
a
Sensitivity was quantitated as described in Materials and Methods for two
different concentrations of maltose. Results from a typical experiment are shown.
VOL. 177, 1995 NlpE OVERPRODUCTION SUPPRESSES PERIPLASMIC LacZ 4221
suspect that the toxicity results from the aborted attempt by the
cell to target the mutant precursor protein to the outer mem-
brane prior to its release from the cytoplasmic membrane. The
cpxA* mutations suppress this growth defect (5a). As expected,
overexpression of NlpE suppresses the defects of lamBA23D
strains. We conclude from all of these results that high levels of
NlpE activate the CpxA kinase. This, in turn, leads to increased
levels of CpxR-phosphate, which activates degP transcription.
When expressed at high levels, a fraction of the NlpE mol-
ecules may fail to fold and/or assemble in the outer membrane
correctly. Indeed, the fractionation experiments shown in Fig.
4 support this view. Accordingly, we have considered the pos-
sibility that CpxA senses overexpressed NlpE as ‘‘junk.’’ We
think this explanation unlikely because overexpression of other
outer membrane proteins, even other lipoproteins, does not
cause Cpx-mediated degP induction (6), nor does it cause sup-
pression. This suggests a normal interaction between NlpE and
CpxA. This interaction could be direct; alternatively, NlpE may
perform some function that CpxA monitors.
To address the critical question of NlpE function, we have
constructed a chromosomal nlpE insertion mutation. Although
some nlpE sequences remain in this construct, we think it likely
that the mutation abolishes NlpE function. Careful examina-
tion of the nlpE mutant strain did not reveal any useful phe-
notypes. In the accompanying paper, Gupta et al. report that
our nlpE null mutant exhibits an increased sensitivity to cop-
per, and sequence analysis reveals a potential heavy-metal
binding site in NlpE (11). These data could indicate a role for
NlpE in copper homeostasis. Alternatively, it may be that the
copper sensitivity observed in the nlpE mutant reflects the
inability of the cell to mount an appropriate stress response.
NlpE may function to sense outer membrane stress and
communicate this information directly to CpxA. Such a model
is attractive because it seems likely that cells have a mechanism
to monitor outer membrane composition. However, a direct
test of this model requires a more complete understanding of
the natural signals that activate the Cpx pathway. At present,
the only known stimulus is overproduction of NlpE, and, ob-
viously, this signal cannot be used to test the involvement of
NlpE.
Examination of the NlpE sequence reveals a serine protease
inhibitor motif starting at amino acid 99 of the mature protein
(Fig. 4A). Although the location of this motif in NlpE is some-
what unusual (15), it is tempting to speculate that it may be
important. It is possible that NlpE functions normally to reg-
ulate the activity of DegP and other periplasmic serine pro-
teases (16). This model is especially attractive because it offers
a plausible explanation for the dosage dependence of NlpE for
suppression of LamB-LacZ-PhoA. This observation is para-
doxical, because high levels of NlpE induce DegP synthesis (6)
but no suppression is observed (Table 1). At moderate levels of
NlpE overexpression, increased amounts of DegP produced by
Cpx-mediated induction are sufficient to degrade the toxic
fusion protein. However, high levels of NlpE may specifically
interfere with DegP activity and/or the activity of the other
factors (see below) that contribute to suppression. Tests of this
model require a more careful examination of this motif.
An alternative explanation for the dosage dependence of
NlpE for suppression posits that very high levels of NlpE sim-
ply saturate the enzymatic activity of DegP. This nonspecific
saturation could also explain why overexpression of other
outer membrane proteins, which induce degP by a different
mechanism, fails to suppress. Here again, DegP may be effec-
tively titrated by the high levels of outer membrane protein.
We do not at present understand why induction of degP by
the Cpx system allows suppression of LamB-LacZ-PhoA but
similar degP induction by overexpression of ompF does not.
Here again, it is possible that high levels of outer membrane
proteins saturate the enzymatic activity of DegP. Perhaps the
cell cannot deal effectively with the stress caused by overpro-
ducing both an outer membrane protein and LamB-LacZ-
PhoA. On the other hand, since degP is controlled by two
different signaling pathways (6), it may be that the two regulons
contain distinct but overlapping sets of genes. Other genes in
the Cpx regulon may contribute substantially to suppression.
As noted above, tests of epistasis demonstrate the impor-
tance of DegP for Cpx-mediated suppression of LamB-LacZ-
PhoA. However, we think it likely that the Cpx two-component
system controls expression of other important genes as well.
Significant suppression by NlpE is still observed in strains that
lack DegP but not in strains that lack CpxA/R (Table 2). In
suppressor strains that lack DegP, the LamB-LacZ-PhoA fu-
sion protein is stabilized but is less toxic than are comparable
levels of fusion protein in wild-type strains. This strongly sug-
gests that the Cpx pathway regulates the synthesis of some
other factors that can combat this periplasmic toxicity by a
mechanism that does not involve degradation. Identification of
these genes and their products may provide important infor-
mation regarding this periplasmic stress response and, per-
haps, the mechanisms employed to fold and target noncyto-
plasmic proteins in general.
ACKNOWLEDGMENTS
We thank Michael Syvanen for providing strains and plasmids used
in this study. We greatly appreciate Ann Flower for help with the
protocol for isolation and fractionation of membranes. Many thanks to
Ken Rudd for suggesting names for the genes identified in this study.
As always, this work would not have been possible without help from
the entire Silhavy laboratory.
This work was supported by a National Institutes of Health grant to
T.J.S. W.B.S., C.L.C., and P.N.D. were supported by a Public Health
Service training grant. W.B.S. was also supported by a Princeton Uni-
versity A-R award.
REFERENCES
1. Bassford, P. J., T. J. Silhavy, and J. R. Beckwith. 1979. Use of gene fusions
to study secretion of maltose-binding protein into Escherichia coli periplasm.
J. Bacteriol. 139:19–31.
2. Bieker, K. L., G. J. Phillips, and T. J. Silhavy. 1990. The sec and prl genes of
Escherichia coli. J. Bioenerg. Biomembr. 22:291–310.
3. Bolivar, F., R. L. Rodriguez, P. J. Green, M. C. Betlach, H. L. Heyneker,
H. W. Boyd, J. H. Crosa, and S. Falkow. 1977. Construction and character-
ization of new cloning vehicles. II. A multipurpose cloning system. Gene
2:95–113.
4. Carlson, J. H., and T. J. Silhavy. 1993. Signal sequence processing is re-
quired for the assembly of LamB trimers in the outer membrane of Esche-
richia coli. J. Bacteriol. 175:3327–3334.
5. Casadaban, M. J. 1976. Transposition and fusion of the lac genes to selected
promoters in Escherichia coli using bacteriophage lambda and mu. J. Mol.
Biol. 104:541–555.
5a.Cosma, C. L., P. N. Danese, J. H. Carlson, T. J. Silhavy, and W. B. Snyder.
Activation of the Cpx signal transduction pathway of Escherichia coli sup-
presses the toxicity conferred by certain envelope associated stresses. Sub-
mitted for publication.
6. Danese, P. N., W. B. Snyder, C. L. Cosma, L. J. B. Davis, and T. J. Silhavy.
1995. The Cpx two-component signal transduction pathway of Escherichia
coli regulates transcription of the gene specifying the stress-inducible
periplasmic protease, DegP. Genes Dev. 9:387–398.
7. Dong, J., S. Iuchi, H.-S. Kwan, Z. Lu, and E. C. C. Lin. 1993. The deduced
amino-acid sequence of the cloned cpxR gene suggests the protein is the
cognate regulator for the membrane sensor, CpxA, in a two-component
signal transduction system of Escherichia coli. Gene 136:227–230.
8. Emr, S. D., and T. J. Silhavy. 1980. Mutations affecting localization of an
Escherichia coli outer membrane protein, the bacteriophage lreceptor. J.
Mol. Biol. 141:63–90.
9. Fellay, R., J. Frey, and H. Krish. 1987. Interposon mutagenesis of soil and
water bacteria: a family of DNA fragments designed for in vitro insertional
mutagenesis of Gram-negative bacteria. Gene 52:147–154.
4222 SNYDER ET AL. J. BACTERIOL.
10. Genetics Computer Group. 1994. Motifs analysis. Program manual for the
Wisconsin Package, version 8, September 1994. Genetics Computer Group,
Madison, Wis.
11. Gupta, S. D., B. T. O. Lee, J. Camakaris, and H. C. Wu. 1995. Identification
of cutC and cutF (nlpE) genes involved in copper tolerance in Escherichia
coli. J. Bacteriol. 177:4207–4215.
12. Hamilton, C. M., M. Aldea, B. K. Washburn, P. Babitzke, and S. R. Kushner.
1989. New method for generating deletions and gene replacements in Esch-
erichia coli. J. Bacteriol. 171:4617–4622.
13. Hayashi, S., and H. C. Wu. 1990. Lipoproteins in bacteria. J. Bioenerg.
Biomembr. 22:451–471.
14. Kohara, Y., K. Akiyama, and K. Isono. 1987. The physical map of the whole
E. coli chromosome: application of a new strategy for rapid analysis and
sorting of a large genomic library. Cell 50:495–508.
15. Laskowski, M., and I. Kato. 1980. Protein inhibitors of proteinases. Annu.
Rev. Biochem. 49:593–626.
16. Lazdunski, A. M. 1989. Peptidases and proteases of Escherichia coli and
Salmonella typhimurium. FEMS Microbiol. Rev. 63:265–276.
17. Lipinska, B., M. Zylicz, and C. Georgopoulos. 1990. The HtrA (DegP)
protein, essential for Escherichia coli survival at high temperatures, is an
endopeptidase. J. Bacteriol. 172:1791–1797.
18. McEwen, J., and P. M. Silverman. 1980. Chromosomal mutations of Esch-
erichia coli K-12 that alter the expression of conjugative plasmid function.
Proc. Natl. Acad. Sci. USA 77:513–517.
19. Mecsas, J., P. E. Rouviere, J. W. Erickson, T. J. Donohue, and C. A. Gross.
1993. The activity of s
E
,anEscherichia coli heat-inducible s-factor, is mod-
ulated by expression of outer membrane proteins. Genes Dev. 7:2618–2628.
20. Osborne, R. S., and T. J. Silhavy. 1993. PrlA suppressor mutations cluster in
regions corresponding to three distinct topological domains. EMBO J. 12:
3391–3398.
21. Rudd, K. E. 1992. Alignment of E. coli DNA sequences to a revised, inte-
grated genomic restriction map, p. 2.3–2.42. In J. Miller (ed.), A short course
in bacterial genetics: a laboratory manual and handbook for Escherichia coli
and related bacteria. Cold Spring Harbor Laboratory Press, Cold Spring
Harbor, N.Y.
22. Russo, F. D., J. M. Slauch, and T. J. Silhavy. 1993. Mutations that affect
separate functions of OmpR, the phosphorylated regulator of porin tran-
scription in Escherichia coli. J. Mol. Biol. 231:261–273.
23. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a
laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold
Spring Harbor, N.Y.
24. Schatz, P. J., and J. Beckwith. 1990. Genetic analysis of protein export in
Escherichia coli. Annu. Rev. Genet. 24:215–248.
25. Silhavy, T. J., M. L. Berman, and L. W. Enquist. 1984. Experiments with
gene fusions. Cold Spring Harbor Laboratory Press, Cold Spring Harbor,
N.Y.
26. Silhavy, T. J., H. A. Shuman, J. Beckwith, and M. Schwartz. 1977. Use of
gene fusions to study outer membrane protein localization in Escherichia
coli. Proc. Natl. Acad. Sci. USA 74:5411–5415.
27. Snyder, W. B., and T. J. Silhavy. 1995. b-Galactosidase is inactivated by
intermolecular disulfide bonds and is toxic when secreted to the periplasm of
Escherichia coli. J. Bacteriol. 177:953–963.
28. Stader, J., S. A. Benson, and T. J. Silhavy. 1986. Kinetic analysis of lamB
mutants suggests the signal sequence plays multiple roles in protein export.
J. Biol. Chem. 261:15075–15080.
29. Stock, J. B., A. M. Stock, and J. M. Mottonen. 1990. Signal transduction in
bacteria. Nature (London) 344:395–400.
30. Strauch, K. L., K. Johnson, and J. Beckwith. 1989. Characterization of degP,
a gene required for proteolysis in the cell envelope and essential for growth
of Escherichia coli at high temperature. J. Bacteriol. 171:2689–2696.
31. Unger, B., G. Klock, and W. Hillen. 1984. Nucleotide sequence of the
repressor gene of the RA1 tetracycline resistance determinant: structural
and functional comparison with three related repressor genes. Nucleic Acids
Res. 12:7693–7703.
32. Weber, R. F., and P. J. Silverman. 1988. The Cpx proteins of Escherichia coli
K-12: structure of the CpxA polypeptide as an inner membrane component.
J. Mol. Biol. 203:467–476.
33. Zhou, Z., and M. Syvanen. 1990. Identification and sequence of the drpA
gene from Escherichia coli. J. Bacteriol. 172:281–286.
VOL. 177, 1995 NlpE OVERPRODUCTION SUPPRESSES PERIPLASMIC LacZ 4223
