(Received for publication, October 29, 1996, and in revised form, April 1, 1997)
From the Amino acid exchanges in the Alcaligenes
eutrophus nickel permease (HoxN) were constructed by
site-directed mutagenesis, and their effects on nickel ion uptake were
investigated. Mutant hoxN alleles were expressed in
Escherichia coli, and activity of the altered permeases was
examined via a recently described physiological assay (Wolfram, L.,
Friedrich, B., and Eitinger, T. (1995) J. Bacteriol. 177, 1840-1843). Replacement of Cys-37, Cys-256, or Cys-318 by alanine did
not severely affect nickel ion uptake. This activity of a C331A mutant
was diminished by 60%, and a similar phenotype was obtained with a
cysteine-less mutant harboring four Cys to Ala exchanges. Alterations
in a histidine-containing sequence motif (His-62, Asp-67, His-68),
which is conserved in microbial nickel transport proteins, strongly
affected or completely abolished transport activity in the E. coli system. The analysis of HoxN alkaline phosphatase fusion
proteins implied that His-62, Asp-67, and His-68 exchanges did not
interfere with overall membrane topology or stability of the nickel
permease. These mutations were reintroduced into the A. eutrophus wild-type strain. Analyses of the resulting HoxN
mutants indicated that exchanges in the histidine motif led to a
clearly decreased affinity of the permease for nickel ion.
Alcaligenes eutrophus H16, a member of the Uptake of nickel ion in sufficient amounts is a prerequisite for the
synthesis of nickel-containing enzymes. Microbial nickel uptake is
mediated by nonspecific transport systems for divalent cations and by
specific systems with a high affinity for Ni2+. In natural
environments the concentration of Ni2+ is generally very
low compared with other divalent cations like Mg2+. Since
nonspecific nickel transport is competitively inhibited by a number of
divalent cations, this mode of uptake is not suited for meeting the
cellular nickel requirements (reviewed in Refs. 2 and 3). Molecular
analyses of Ni2+-specific transport systems of a few
bacteria containing [NiFe] hydrogenases and/or ureases showed that
two different types of membrane transporters are responsible for high
affinity uptake of the transition metal (reviewed in Ref. 3). The best
studied Ni2+ transporters, the Nik system of
Escherichia coli (4) and the HoxN permease of A. eutrophus (5, 6), are the prototype of these two classes. Nik
belongs to the ATP binding cassette family of transporters. It consists
of five subunits: a periplasmic binding protein (NikA), two integral
membrane subunits (NikB, NikC), and two components containing ATP
binding motifs (NikD, NikE). There is no evidence that HoxN requires
additional proteins for its function as a permease. This suggests that
the two types of transporters employ different uptake mechanisms. HoxN
is an integral membrane protein and was until recently thought to
contain seven transmembrane segments (6, 7). The topological model was
based on the construction and analysis of a number of fusions to
alkaline phosphatase and
HoxN homologs have been found in Bradyrhizobium japonicum
(8), a hydrogenase-producing soil bacterium; in the thermophilic Bacillus strain TB-90 (9), which contains a urease; and in Helicobacter pylori (10), a human pathogen that is dependent on a highly active urease for initial colonization of the gastric mucosa. It is still not known how these transporters discriminate among
divalent cations. Biochemical analyses of the HoxN-type permeases are
difficult because these systems have a high affinity for
Ni2+ (KT of 10-20 nM)
but a very low capacity (reviewed in Ref. 3). Therefore, we developed a
physiological assay for HoxN-mediated Ni2+ uptake (7) to
carry out a detailed analysis of the structure-function relationship.
In the present report we describe the consequences of amino acid
replacements in HoxN. We chose the cysteine residues because of their
hypothetical role in metal binding. The main focus of this study,
however, is on a conserved histidine motif (His-Xaa4-Asp-His) found in HoxN-type permeases as well as
in NikC, an integral membrane component of the E. coli
nickel transport system (7).
Analytical grade chemicals were obtained from
Merck, Sigma, and Fluka. 63NiCl2 (30 TBq/mol)
and [35S]dATP E. coli CC118
(11) was used as the host for recombinant plasmids. Plasmids pCH231-Sm
harboring the A. eutrophus gene hoxN, pCH231-P47,
which contains an insertion of the transposon TnphoA in the
48th codon of hoxN, and pCH231-P291 encoding a fusion of E. coli alkaline phosphatase (PhoA) to HoxN residue 291 were
described recently (6, 7). Plasmid pKAU17 (12) containing the urease operon of Klebsiella aerogenes was a gift from R. P.
Hausinger (Michigan State University, East Lansing). hoxN
mutations were transferred to A. eutrophus H16 (DSM 428, ATCC 17699) by conjugation using the donor E. coli S17-1
(13) and the counterselectable suicide vector pLO2 as described by Lenz
et al. (14).
Mutagenesis
was carried out by the PCR-based method of Chen and Przybyla (15),
which in most cases requires two rounds of PCR but only a single
mutagenic oligonucleotide primer. The mutagenic oligonucleotides used
in this study are listed in Table I. Taq DNA polymerase was
used for DNA amplification. To avoid unwanted mutations generated by
Taq-catalyzed untemplated addition of nucleotides, fragments
of the first PCR round were treated with T4 DNA polymerase, which
removes 3 Table I.
Mutagenic oligonucleotides used for amino acid replacements
Humboldt-Universität zu Berlin,
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
division
of the proteobacteria, can utilize various organic compounds or
molecular hydrogen as energy sources. Hydrogen oxidation is catalyzed
by two metalloenzymes: a cytoplasmic NAD+-reducing
hydrogenase and a membrane-bound, electron transport-coupled hydrogenase. Both enzymes belong to the family of [NiFe] hydrogenases (reviewed in Ref. 1). A nickel-dependent urease allows the organism to grow on urea as a nitrogen source.
-galactosidase (6). Due to an error in the
originally published hoxN nucleotide sequence (5), the
C-terminal part of the protein was neglected in the topological study.
Additional fusions favor the eight-helix model presented in Fig.
1.
Fig. 1.
Revised topological model of the A. eutrophus nickel permease. Residues that were subjected to
site-directed mutagenesis are shown in solid boxes. See text
for details.
[View Larger Version of this Image (20K GIF file)]
Materials
S1 were from
Amersham. Zinsser Aquasafe 300 Plus was used as the mixture for liquid
scintillation analyses in a Packard 1600 TR counter. Biochemicals were
obtained from Boehringer Mannheim, New England Biolabs, Amersham, and
Life Technologies. PhoA hybrid proteins were detected in Western
immunoblots using a rabbit anti-PhoA antibody (5 Prime
3 Prime,
Inc., Boulder, CO) and an alkaline phosphatase-coupled goat anti-rabbit
antibody (Jackson ImmunoResearch). Nitrocellulose blotting membranes
were from Schleicher & Schüll. For direct sequencing of
PCR-amplified DNA, samples were purified on Qiagen spin columns.
overhangs. The products of the second PCR round were
digested with appropriate restriction endonucleases, purified on
agarose gels, and used to replace the respective fragments of plasmid
pCH231-Sm. The complete sequences of synthetic DNAs were verified by
the dideoxy chain termination method of Sanger et al. (16).
Modified hoxN segments were introduced into a derivative of
the A. eutrophus wild-type strain
H16,2 which harbors a large in-frame
deletion in hoxN covering 768 bp. The allelic exchange
procedure took advantage of the conditionally lethal sacB
gene on vector pLO2 (14). Sucrose-resistant colonies were screened for
full-length hoxN by PCR amplification. As described by
Bernhard et al. (17), 1 ml of a fresh overnight culture
grown in nutrient broth was concentrated 2-fold in distilled water and boiled for 5 min. 10 µl of this crude lysate was used for PCR. hoxN point mutations in A. eutrophus were
confirmed by sequence determination of the amplified fragments. Genomic
rearrangements adjacent to the recombination sites were excluded by PCR
amplification of a segment spanning hoxN and the 5
and 3
neighboring regions.
Mutant
Orientationa
Oligonucleotide
(5
3
)b
Codon change
H62A
Antisense
CGTCTACCGCNNHGCGCAACCCAAGGCC
CAC
GCG
H62I
Antisense
CGTCTACCGCNNHGCGCAACCCAAGGCC
CAC
ATC
H62S
Antisense
CGTCTACCGCNNHGCGCAACCCAAGGCC
CAC
AGC
D67E
Antisense
TTGCCGCGAGATGCTCTGCGTCTACCGCGTGG
GAT
GAG
D67N
Antisense
TTGCCGCGAGATGATTTGCGTCTACCGCGTGG
GAT
AAT
H68I
Antisense
TTGCCGCGAGAATATCTGCGTCTACCGCGTGG
CAT
ATT
H68S
Antisense
TTGCCGCGAGACTATCTGCGTCTACCGCGTGG
CAT
AGT
C37A
Sensea
GGCTCGCCGCGCTCGCCGCG
TGC
GCC
C256A
Sense
GACACCATCGATAGCATCCTGATGGCCGGCGCCTATGCGTGG
TGT
GCC
C318A
Sense
GAATTTCGCCCAATTGGGCTTCGTC
TGT
GCC
C331A
Sense
TACCGTGGCCTGGGTCGTGTCGATCG
TGT
GCC
a
Orientation of the oligonucleotide with respect to the
orientation of the hoxN reading frame.
b
N stands for a mixture of A, C, G, and T. A mixture of A, C,
and T is indicated by H. Altered bases are shown in boldface italics.
Nickel accumulation of E. coli CC118 harboring plasmid pCH231-Sm or derivatives was tested as described recently (7). The cellular content of 63Ni is expressed as picomoles per milligram of protein. Urease (EC 3.5.1.5) activity of strain CC118 containing pKAU17 and pCH231-Sm or derivatives was determined upon growth in the presence of 500 nM NiCl2 by quantitating the rate of ammonium ion released from urea (7). Alkaline phosphatase (EC 3.1.3.1) activity was assayed colorimetrically using 4-nitrophenyl phosphate as the substrate (6). The A. eutrophus soluble hydrogenase (SH; EC 1.12.1.2) was measured with permeabilized cells by monitoring H2-dependent reduction of NAD+ as described by Friedrich et al. (18). Specific enzyme activities are given in units per milligram of protein. One unit of activity is defined as the amount of enzyme required to consume 1 µmol of substrate per min. Protein was estimated by the method of Lowry et al. (19).
Western Immunoblot AnalysisCells were grown in Luria-Bertani broth, harvested, and resuspended in buffer containing 3% (w/v) SDS. 30 µg of protein was separated by SDS-PAGE (7.5% polyacrylamide) (20) and electroblotted onto nitrocellulose membranes. Immunoblots were developed as described recently (6).
In the course of
site-directed mutagenesis we identified an error in hoxN. An
additional G is present between bases 1,102 and 1,103 of the published
sequence (5) giving an open reading frame of 1,053 bp. Thus, the
deduced polypeptide consists of 351 rather than 301 amino acid residues
and has a molecular mass of 38.9 kDa. Computer analyses of the
corrected nucleotide sequence predicted a membrane-spanning segment in
the new C-terminal part of the HoxN protein. The previous topological
analysis, which was confined to the first 301 amino acids, identified
seven transmembrane helices (6). Data for additional alkaline
phosphatase and -galactosidase fusions to residues 300 and 351 of
the permease, respectively (not shown), support the assumption of an
eighth membrane-spanning helix in the C-terminal part of the protein.
The revised topological model is shown in Fig. 1.
Mutations in
single cysteine codons of hoxN were generated by PCR using
the mutagenic oligonucleotide primers listed in Table I
and plasmid pCH231-Sm as template. This procedure led to the isolation
of permease mutants with Cys-37, Cys-256, Cys-318, or Cys-331 replaced
by Ala. Double and triple mutants and a quadruple mutant with two,
three, and four Cys to Ala exchanges, respectively, were obtained as
follows. C256A/C318A, C256A/C331A, and C318A/C331A mutants were
generated via PCR using single mutation plasmids as templates. The
C37A/C256A, C37A/C318A, and C37A/C331A derivatives were constructed by
insertion of a 303-bp ClaI/Bsu36I fragment with
the respective 3 mutation into a plasmid carrying the C37A mutation.
Both strategies were used to generate the four triple mutants.
Introduction of the 303-bp fragment corresponding to the
C256A/C318A/C331A triple mutant into the C37A plasmid yielded the
cysteine-less quadruple mutant. His-62 exchanges were produced via PCR
using an oligonucleotide primer with a "three-base wobble" (Table
I) at the relevant position. This method permitted the isolation of
mutants with H62A, H62I, and H62S replacements. Additional His-62
mutants, obtained by this strategy, were not further analyzed since
spontaneous second site mutations had occurred during PCR amplification. Asp-67 and His-68 mutants were constructed using the
mutagenic primers listed in Table I. The H62I/H68I double mutant was
obtained by introduction of a 96-bp AccI/MscI
fragment containing the H68I mutation into a plasmid harboring the H62I mutation.
To investigate the role of the residues Cys-37, Cys-256, Cys-318, and Cys-331 in Ni2+ uptake we assayed Ni2+ accumulation and urease activity of E. coli CC118 producing altered HoxN proteins with single or multiple Cys replacements. The results are shown in Table II. E. coli transformants producing the wild-type A. eutrophus nickel permease accumulated 34 pmol of Ni2+ per mg of protein and displayed a urease activity of 581 milliunits per mg of protein. In comparison, the control strain with an inactivated hoxN accumulated 1.7 pmol of Ni2+ per mg of protein and showed a urease activity of 27 milliunits per mg of protein. Both Ni2+ accumulation and urease activities were unaffected by the C37A exchange. The C256A mutant harboring a replacement within transmembrane helix VI (Fig. 1) showed 80% of the wild-type level of Ni2+ accumulation and 60% of the urease activity suggesting that Cys-256 is not essential for metal ion transport. A mutation resulting in the replacement of Cys-318 by Ala led to a slight but reproducible increase of Ni2+ uptake. Cys-318 is located within the fourth periplasmic loop connecting helices VII and VIII (Fig. 1). It is tempting to speculate that this residue has metal binding capacity but is not part of the substrate recognition domain, and thus, its replacement may increase Ni2+ transport across the membrane. Among the Cys single mutants, the C331A exchange gave the strongest response resulting in a 50-70% decrease in Ni2+ uptake. The analysis of Cys double and triple mutants (data not shown) did not add to the information obtained with the single replacements. None of the mutants harboring the C331A exchange had more than 50% of the Ni2+ uptake activity compared with the wild-type HoxN. This also applies to the cysteine-less quadruple mutant that shows only 30% of the normal Ni2+ accumulation and urease activities (Table II).
|
In the course of recently reported sequence alignments of nickel transport proteins (7) we identified the conserved motif His-Xaa4-Asp-His that resembles the Ni(II)-binding site of human serum albumin and is located within the second transmembrane helix of the HoxN-type permeases (3) (Fig. 1). Since this was the first hint to a substrate-binding domain in these proteins we investigated the significance of this signature for the activity of the A. eutrophus HoxN. His-62 was replaced by Ala, Ile, and Ser; Asp-67 by Asn and Glu; and His-68 by Ile and Ser. The effect on Ni2+ accumulation and urease-enhancing activity was monitored in E. coli. The replacements had clear-cut effects. Exchanging His-62 to Ile abolished Ni2+ uptake almost completely, and the H62A mutant showed only marginal activity, while the H62S mutant contained residual activity of approximately 20% (Table II). These results are compatible with the notion that the nature of the amino acid side chain at position 62 of the permease is critical. On the other hand, all Asp-67 and His-68 mutants were totally devoid of activity (Table II).
Expression of Permeases with His-62, Asp-67, and His-68 ExchangesTo exclude the possibility that diminished transport
activity of the mutants is due to altered gene expression, defective insertion of the permease into the membrane, or enhanced proteolysis, we studied the influence of the mutations on HoxN levels by Western immunoblot analysis. Since an antibody against HoxN is not available, we took advantage of a previously described fusion of E. coli PhoA to HoxN (6). In this construct (P291) PhoA is fused to amino acid 291 of HoxN and faces the periplasm (Fig. 2).
P291 derivatives with exchanges at position 62, 67, and 68 were
generated by introducing a 1.3-kilobase XhoI/MscI
fragment containing the respective mutation into the parent P291 gene
fusion. PhoA activities of P291 derivatives were measured, and the
amount of the fusion proteins in the E. coli CC118 membrane
was determined by immunoblotting using an anti-PhoA antibody. The
results are shown in Fig. 2. Comparable quantities of P291 were
detected irrespective of the His-62, Asp-67, and His-68 exchanges, and
each fusion displayed PhoA activity. A remarkable decrease of this
activity was observed for the D67N and H68I mutants. Nevertheless,
these results indicate that the altered permeases are present in
similar amounts and correctly oriented in the membrane.
Influence of Replacements in the Histidine Motif on Hydrogenase Activity of A. eutrophus
To assay the mutant permeases in the homologous background we reintroduced the amino acid exchanges into the A. eutrophus wild-type strain. Ni2+ transport experiments assaying the activity of HoxN in A. eutrophus or in recombinant E. coli do not yield reliable results since nonspecific binding of Ni2+ to the cell envelope and interference by the Mg2+ transport systems hinder kinetic analyses. Therefore, we monitored the effect of the replacements on the activity of the SH, a nickel-containing enzyme of A. eutrophus. The results are summarized in Table III. Surprisingly, the His-62 mutants were not impaired in SH activity when the cells were grown in a mineral salts medium supplemented with 100 nM NiCl2. Under these conditions, however, SH activity of a strain with a deletion in hoxN was clearly reduced. These results indicate that His-62 exchanges have no effect on the apparent activity of the nickel permease under these circumstances. In an additional experiment, the SH activity of mutants grown in mineral medium that was not supplemented with NiCl2 was measured. Under Ni2+ starvation, only 10% of SH activity was detectable in the wild-type strain and in the H62A and H62S mutants. SH activity in the H62I strain was less than 3% and was almost undetectable in the hoxN deletion mutant. These results indicate that His-62 has a significant effect on the affinity of the permease for the Ni2+ ion. Experimental support for this conclusion was obtained by growing the cells in a medium containing 100 nM NiCl2 and 10 µM nitrilotriacetate (NTA), a well known Ni2+ chelator. In this case Ni2+ is present almost exclusively in the form of NTA complexes. Addition of the metal-complexing agent did not affect SH activity in the wild-type strain but completely abolished the SH activity in the strain lacking HoxN. SH activities of the H62A, H62I, and H62S mutants were 65, 40, and 75%, respectively, compared with the wild type. In the presence of 50 µM NTA, the SH activities of the mutants were reduced to background levels. The Asp-67 and His-68 mutations as well as the H62I/H68I double mutation in hoxN had a dramatic effect. SH activities of these strains were almost undetectable under any conditions of nickel limitation tested. These results confirm that wild-type HoxN mediates maximal Ni2+ uptake even at extremely low substrate concentrations. Amino acid exchanges at positions 67 and 68 apparently inactivate the permease whereas exchanges at position 62 interfere with the affinity of the transporter for Ni2+ ion.
|
Ni2+-specific transport systems have been shown to play a key role in the synthesis of bacterial NiFe hydrogenases and ureases that are involved in energy and nitrogen metabolism of the respective species (reviewed in Ref. 3). In H. pylori, a human pathogen that is dependent on a highly active urease, Ni2+ transport is essential for pathogenesis (10). So far two different types of high affinity Ni2+ transporters have been identified. The Nik system, a permease belonging to the ATP binding cassette family, is to date unique to E. coli (4). Homologous one-component nickel permeases have been found in A. eutrophus (HoxN (5-7)), B. japonicum (HupN (8)), and H. pylori (NixA (10)). These systems show a common topology with eight membrane-spanning segments (reviewed in Ref. 3). The UreH protein of thermophilic Bacillus strain TB-90 (9), implicated in Ni2+ transport and clearly related to the one-component systems, probably contains less than eight transmembrane helices. An important feature of these Ni2+ transport systems is their ion selectivity allowing sufficient Ni2+ uptake even in environments where only traces of the transition metal ion are available. In a very recent study (23), a novel transporter (NhlF) mediating Co2+ uptake in the Gram-positive bacterium Rhodococcus rhodochrous J1 was described. This organism produces two types of cobalt-containing nitrile hydratases. Interestingly, NhlF is homologous to the one-component nickel permeases. Alignments revealed more than 35% amino acid sequence identity between NhlF and HoxN, HupN, and NixA (23). Nevertheless, NhlF does not catalyze Ni2+ transport. Co2+ transport, however, was inhibited by a 500-fold excess of Ni2+ ion, while other divalent cations had no effect (23). Thus, learning how these systems discriminate among divalent cations is essential for an understanding of the transport mechanism. In the E. coli Nik system, ion selectivity may be inherent to the periplasmic binding protein. In the case of the one-component permeases, specificity is an intrinsic property of the membrane proteins.
The present study represents a first attempt to gain insight into the structure-function relationship of the nickel permease (HoxN) of A. eutrophus. Using site-directed mutagenesis we introduced amino acid replacements in a putative nickel binding motif and at the positions occupied by Cys residues. These Cys residues are not conserved in the one-component Ni2+ transport systems, but nevertheless thiol groups are candidates for metal recognition. Evidence obtained with the Cys single, double, triple, and quadruple mutants suggests that Cys-331, located within transmembrane helix VIII, is the only Cys residue that could be directly involved in substrate recognition. Although not leading to a detailed insight into the transport mechanism, the availability of the Cys-less quadruple mutant is helpful for further studies such as cysteine-scanning mutagenesis. The latter technique in connection with excimer fluorescence labeling or permease probing with impermeant SH-reactive agents was successfully applied in studies on the three-dimensional structure of the lactose permease (21) and the mechanism of the glucose 6-phosphate/phosphate antiporter of E. coli (22). Cys-free variants are a prerequisite for this approach.
Replacements of His-62, Asp-67, and His-68, which are conserved in bacterial nickel permeases (3, 7), had clear-cut effects in the physiological assay system. Mutant permeases were dramatically impaired in Ni2+ uptake when cells were grown in complex medium containing 1% tryptically digested protein and 0.5% yeast extract. This medium has a strong metal-chelating capacity. Under these conditions, altered permeases harboring hydrophobic residues at position 62 were almost completely unable to mediate Ni2+ uptake. A hydrophilic Ser residue at this position, however, correlated with a residual activity. All Asp-67 and His-68 exchanges inactivated the permease. These results suggest that all three residues of the conserved motif participate in substrate recognition. Introduction of the mutations into A. eutrophus and measurement of the effects on Ni2+ uptake confirmed this assumption. In the presence of free Ni2+ in a mineral medium, the permeases altered at position 62 were capable of metal ion uptake. At an extremely low Ni2+ concentration or in the presence of Ni2+ complexed with nitrilotriacetate, however, transport activity was clearly reduced indicating that the nature of the side chain at position 62 is important for affinity.
Of particular importance for further investigations is the finding published by Komeda et al. (23) that the Co2+ transporter of R. rhodochrous reveals significant sequence similarity with the A. eutrophus HoxN and contains the His-Xaa4-Asp-His motif. The presence of this putative Ni2+ binding motif could explain the inhibition of Co2+ uptake by Ni2+ ion. On the other hand, previous uptake experiments (24) and the analysis of Ni2+ accumulation in the present study (data not shown) indicated that HoxN-mediated Ni2+ transport is only slightly affected in the presence of Co2+ ion. This suggests that HoxN lacks a Co2+-binding site. Although much more work remains to be done, the results presented in this report are a basis for elucidating the principles of ion recognition of metal-specific permeases.
We thank Bärbel Friedrich for generous support, Edward Schwartz and B. Friedrich for critical comments on the manuscript, and Werner Schröder for the synthesis of oligonucleotide primers. We are also indebted to Robert P. Hausinger for stimulating suggestions and to Marion Müller for skillful technical assistance.