|
|
|
|||||||
|
|||||||||
(Received for publication, November 9, 1994; and in revised form, January
23, 1995) From the
Two uniquely paired deoxynucleoside kinases, deoxycytidine
kinase/deoxyadenosine kinase (dCK/dAK) and deoxyguanosine
kinase/deoxyadenosine kinase (dGK/dAK) are required, together with
thymidine kinase (TK), for deoxynucleotide synthesis in Lactobacillus acidophilus R-26. Using polymerase chain
reaction-generated probes based on N-terminal amino acid sequences, we
have cloned tandem genes for 25- and 26-kDa polypeptides, whose derived
amino acid sequences and size correspond to wild-type Lactobacillus enzyme subunits. Expression in Escherichia coli uses a
single endogenous promoter and yields active dGK/dAK (
Lactobacillus acidophilus R-26 requires a source of
deoxyribose (1) and is dependent upon four deoxynucleoside
kinases to supply its DNA precursor nucleotides, having no functional
ribonucleotide reductase(2) . While most bacteria have
thymidine kinase (TK), ( Thus, the
presence of all four kinases as obligatory activities in L.
acidophilus may well be unique among the bacteria. Its thymidine
kinase closely resembles the E. coli enzyme (8) in its
kinetic and regulatory behavior, ( While it
is unclear what advantage is conferred by the separate expression of
two subunits specific for deoxyadenosine, it is known that the total
activities toward deoxyadenosine, deoxycytidine, and deoxyguanosine are
about equal when both of the deoxyadenosine kinases are allosterically
activated by half-saturation of the opposing subunit with deoxycytidine
and deoxyguanosine, respectively(10, 11) . In
addition, each activity is controlled by means of end product
inhibition by its respective deoxynucleoside triphosphate, which
apparently binds to both the nucleoside and triphosphate sites in the
manner of a bisubstrate analog, thereby achieving feedback control in a nonallosteric fashion(9) , as has also been proposed
for human deoxycytidine kinase(12) . So that the mechanisms of
these effects on the enzyme proteins can be studied by chemical and
physical means, it is clearly desirable to clone their genes and to
express greater quantities of these proteins. It is also of great
interest to learn how these genes are organized for expression by the
bacterial genome. Molecular cloning of the
deoxyguanosine/deoxyadenosine kinase has now been achieved and the DNA
sequences determined. Development of a cloning probe was complicated by
the presence of the highly conserved glycine-rich ATP-binding site very
near the N terminus (9, 13) of each subunit, as well
as a high degree of codon degeneracy, but the polymerase chain reaction
was used to provide the necessary selection specificity. The enzyme has
been expressed efficiently in an E. coli host, utilizing the Lactobacillus promoter.
Colony hybridization was performed by established
procedures(17) , but signals were detected nonisotopically by
the Photogene nucleic acid detection system (Life Technologies, Inc.),
as described in the manual.
Asymmetrical PCR
was employed to generate single-stranded enriched DNA (19, 20, 21) so that the probe could be
sequenced directly. Sense and antisense primers were at a 1 to 20 molar
ratio, or vice versa. The lower concentration primer was set at 10
nM so it would be depleted after 10-15 cycles, and
1-10 ng of template DNA were used. Other reaction components and
concentrations were the same as for the regular PCR described above. In
the last cycle, the reaction was extended at 72 °C for 7 min.
Single-strand enriched PCR products were purified on a Qiagen tip-5
column, and DNA sequencing was performed, with the lower concentration
PCR primer as the sequencing primer. Extraction and Purification
of Enzymes from E. coli Clones-Transformed E. coli XL1-Blue cells were grown at 37 °C in LB broth containing 75
µg/ml ampicillin to an A
Because the 3` end of one of the paired genes was missing, a
new library had to be constructed. To be certain of cloning the same
genes, a more specific probe was prepared by PCR. A new sense primer,
5`-ACTAGTTAACGAATAGAAGG-3`, a 20-mer reflecting the short noncoding
sequence between the tandem genes, was used, keeping the original
antisense primer. This new probe was therefore a 117-mer specific for
the beginning of the fragmentary gene. Knowing that the 5` to 3`
orientation of the genes was from the KpnI to XbaI,
the 2.5-kb KpnI genomic fragment that hybridized to the probe
on the Southern blot could be expected to include all of the downstream
gene. Therefore, a partial genomic library was constructed from the
2-4-kb region of a KpnI digest, also in pBluescript, and
transformed into the E. coli XL1-Blue cells. After initially
screening more than a thousand recombinants by colony hybridization
with the new probe, a secondary PCR screening of groups of 20 colonies
was carried out, using the primers employed in constructing the probe.
Then, each colony of the pooled group which gave rise to the expected
117-bp fragment was screened individually by PCR. A positive clone,
GTM-K48, was identified and was found to contain nucleotide sequences
encoding the N-terminal amino acids of dAK and dCK (or dGK).
Figure 1:
Sequence of the dAdo kinase and dGuo
kinase genes from L. acidophilus. Lowercase letters indicate the nucleotide sequence; uppercase letters indicate the amino acid sequence. Promoter elements (-35,
-10, etc.) and ribosome binding sites (S/D) are underlined and indicated as marked. The transcription
terminator is marked by opposing arrows.
Weak ``-10'' and
``-35'' promoter elements precede the dAK gene, with
the sequences TACACT and TTGTTT beginning at nucleotide positions
-39 and -63, respectively. Additional promoter elements
known to be conserved among Gram-positive bacteria (23) are
also found. The first is the ``-45'' A cluster (AAAAA)
beginning at nucleotide -72; the second is the T sequence at
nucleotide -75; and finally the TG sequence at nucleotides
-44. The sequence GAAAGA, 6 bases upstream from the translation
start codon of the dAK gene, is a variant of the consensus
ribosome-binding site. Preceding the second gene (the dGK gene),
sequences homologous to the ``-10'' and
``-35'' promoter elements are positioned at 141, 259,
331, and 355 bases upstream from the initiation codon of the second
gene and overlap the coding region of the first gene. However, since
none of the conserved Gram-positive promoter elements precedes the
second ORF and also since possible promoter elements are not close to
the gene, it is likely that only one promoter functions in the
transcription of both genes. There is also a putative ribosome-binding
site 6 bases upstream from the initiation codon of the second gene,
with the sequence GAAGGA. Just 4 bases downstream from the
translational termination codon (TAA) of the second gene, a sequence
having the characteristics of a transcription terminator (24) is seen. A GC-rich region of dyad symmetry, AAAACTGCGGTCCA
and TGGATCGCAGTTTT (marked by arrows in Fig. 1), which
in the transcript may form a stem and loop hairpin structure, is
followed by a stretch of thymidines (positions 1531-1536),
another feature of transcription terminators. No such characteristic
sequence is observed near the stop codon of the dAdo kinase gene, so it
seems likely that the two subunits are transcribed as a polycistronic
mRNA.
To verify
that this cloned dGK/dAK is similar to the wild-type Lactobacillus dGK/dAK, the enzyme activities were tested for allosteric
interactions and end product inhibition. Like the wild-type enzyme, the
cloned dAK in crude extract was activated about 5-fold by 0.1 mM dGuo, and also was inhibited 90% by 0.5 mM dATP; dGK was
also potently inhibited by 0.5 mM dGTP, with only 5% activity
remaining. However, dCyd has no stimulating effect on dAdo
phosphorylation, again indicating that the dAK cloned is the one
associated with the dGK, not with the dCK subunit. The cloned
dGK/dAK could be purified to apparent homogeneity by dATP-Sepharose
affinity chromatography in exactly the same fashion as the wild-type
enzyme(22) . The specific activities of the pure enzyme were
1000 and 65 units/mg for dGK and dAK, respectively, compared with 2000
and 250 units/mg for the wild-type enzyme. We suspected that these
differences might be due to differences in the processing mechanisms in Lactobacillus versus the E. coli host. Comparing the
inferred N-terminal amino acid sequences with sequences of wild-type
dGK and dAK subunits, obtained previously(9) , reveals two
discrepancies (Fig. 2). The cloned dAK gene has codons for two
additional amino acids (Thr-Val) immediately following the initial
methionine, and the initiating methionine encoded in the cloned dGK
gene is missing from the wild-type dGK protein sequence. While it is
not uncommon to have the initial methionine processed off, the Lactobacillus strain seems to possess some sort of a mechanism
to remove the Thr and Val residues, while leaving the initial
Met.
Figure 2:
Comparisons of N-terminal amino acid
sequences determined for wild-type dGK and dAK subunits (9) with sequences inferred from cloned DNA
sequences.
Preliminary N-terminal sequence analysis was performed on the
purified dimeric expressed protein, anticipating two residues per cycle
if the cloned enzyme were identical to wild type. However, each cycle
yielded only one amino acid residue, in order: Thr, Val, Ile, Val, Leu,
and Ser, exactly as found in the wild-type dGK subunit. The absence of
N-terminal Met may have one of two interpretations. It could have been
processed off of both subunits identically, or these results could mean
that the dAK subunit is blocked when expressed in E. coli, and
that only dGK could be sequenced. The latter seems more likely, since
only about 25% of the total protein appeared to be sequenceable. The
purified cloned dGK/dAK migrated at the same rate on SDS-polyacrylamide
gels as the wild-type dGK/dAK, both at pH 9.5 (the pH in the running
gel of the Laemmli system) and at pH 6.8 in the multiphasic zone
electrophoresis 3328.IV buffer system(25) , consisting of
BisTris/TES and BisTris/Cl buffers, employed by Moos et
al.(26) . In the latter buffer system, the subunits were
resolved into two bands, just as they were with wild-type enzyme (9) (results not shown).
Figure 3:
Homology of derived amino acid sequences
of dAdo kinase (dAK) and dGuo kinase (dGK). 61%
identity in the 215-amino acid overlap (DNAStar AANW). A colon (:)
denotes amino acids that are positively related using the probability
of acceptable mutation matrix, a blank denotes negatively
related amino acids, and a period (.) indicates a neutral relationship.
Consensus regions, sites i, ii, iii, and iv, described in the text, are indicated in bold.
Recent reports of sequences inferred from
several mammalian deoxycytidine kinase clones prompted us to attempt
multiple alignment with human(28) , mouse(29) , and rat
(GenBank
Figure 4:
Multiple alignment of dGK and dAK
sequences with mammalian deoxycytidine kinases. mdCK, mouse; hdCK, human; rdCK, rat. The solid bullet marks residues common to all five sequences; the open bullet marks residues found in all three dCKs plus dGK. Sequence
comparisons were facilitated by the Blast E-mail server at the National
Center for Biotechnology Information at the National Library of
Medicine (43) and the AllAll E-mail server of the Computational
Biochemistry Research Group at the ETH in Zürich,
Switzerland.
Finally, it is
interesting to note that one other region, residues 153-161 of
the dGK subunit (site iv), is homologous to the highly
conserved region G-2 of p21
Figure 5:
Comparison of dGK and dAK residues
153-161 with the region G-2 consensus sequence of Ras-like
proteins(31) . Residues in common are in boldface.
dGK/dAK plays an essential role in generating the
deoxyribonucleotide precursors, dGTP and dATP, for DNA metabolism in L. acidophilus R-26. Recently, workers in this laboratory have
purified the paired wild-type dGK/dAK to homogeneity by the successive
application of dCTP-Sepharose and dATP-Sepharose affinity
chromatography and have demonstrated that the enzyme is composed of two
nonidentical subunits of similar molecular mass ( Clone GTM-K48 contains two intact genes,
separated by a 21-bp spacer and a single translation termination codon.
This tandem arrangement, with a functional promoter upstream and a
transcription terminator loop downstream, suggests that a polycistronic
message for the two subunits is transcribed from a new operon of Lactobacillus. The amino acid sequence inferred from the
upstream gene matches the known N-terminal sequences of dAK, and the
downstream gene corresponds to those of either dGK or dCK (after
processing). However, the gene products that are readily expressed in
active form in the E. coli host are specific for dGuo and dAdo
only, not for dCyd. Identities of the two genes as those of dGK/dAK are
further supported by the allosteric interaction and product inhibition
patterns characteristic of dGK/dAK(30) . The cloned enzyme
appears to be identical to the wild-type Lactobacillus enzyme
in virtually every respect. It was purified to homogeneity by
dATP-Sepharose affinity chromatography exactly like the wild-type
enzyme(22) . Also like wild-type enzyme, the pure cloned enzyme
exhibits an apparent subunit mass of These results, taken together with the genetic and kinetic data from
our companion study(30) , clearly demonstrate that the first of
the tandem cloned genes encodes dAK and the second gene encodes dGK.
This tandem arrangement of the genes on the Lactobacillus chromosome, the presence of a functional promoter sequence
preceding the dAK gene, and the apparent transcription terminator loop
following the dGK gene strongly suggest that a polycistronic message is
encoded in the transcription step, and this is a subject for further
investigation. Further study is needed also to learn whether the
apparent deletion of the N-terminal threonine and valine (codons 2 and
3) from wild-type dAK subunits, but with retention of the initiating
methionine, represents some new type of post-transcriptional or
-translational modification occurring in lactobacilli. Given the fact
that the first 19 codons of dAK are identical with those of dGK, it is
difficult to see how such a mechanism could be applied selectively to
dAK. Another surprising observation to be drawn from the inferred
amino acid sequences, is that, unlike mammalian dCKs which require
thiols for activity(33, 34) , neither subunit of the
cloned dGK/dAK contains a cysteine (nor is thiol required for
activity). The absence of cysteine from wild-type enzyme has been
confirmed by amino acid analysis of carboxymethylated protein. ( Codon usages of the dGK/dAK genes are similar to other L.
acidophilus genes in that they appear to have a preference for U
and A in the third base of synonymous codons. This bias against codons
ending in G or C also appears in Lactobacillus helveticus and Lactobacillus lactis, but is less pronounced in other Lactobacillus species(35) . A search of
GenBank The considerably higher sequence homology with mammalian dCKs,
contrasted with TKs, is interesting in the light of their functional
and regulatory similarities. Mammalian dCKs phosphorylate dAdo and dGuo
in addition to dCyd, albeit at a common catalytic site. Also, there are
similarities in the apparent mode of end product inhibition of dGK/dAK
and mammalian dCKs. Whereas thymidine kinase appears to behave like a
classical allosterically inhibited enzyme having cooperative kinetics
and an inhibitor site(8, 42) , there is evidence that
human dCK, at least, is controlled by the nonallosteric multisubstrate
inhibitor mechanism we have postulated for these Lactobacillus enzymes (12) . The longer consensus sequence accompanying
the D/ERS motifs suggests functional importance, which by analogy with
TKs may somehow contribute to deoxynucleoside binding. However, it
seems unlikely to be a specificity determinant, since these larger
sites are nearly identical in the two subunits of dGK/dAK. The
presence of a highly conserved sequence, -YDPTLEDYY- (Fig. 3, site iv) characteristic of Ras-like proteins, is certainly
unexpected in the deoxynucleoside kinases. The threonine of this motif,
found in the dGK but not dAK of our sequences, is critical in
p21
The nucleotide
sequence(s) reported in this paper has been submitted to the
GenBank(TM)/EMBL Data Bank with accession number(s)
U01881[GenBank].
Volume 270,
Number 12,
Issue of March 24, 1995 pp. 6595-6601
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
3% of
extracted protein) closely resembling wild-type dGK/dAK in specificity,
kinetics, heterotropic activation, and end product inhibition.
Alignment of cloned genes reveals 65% identity in their DNA sequences
and 61% identity in derived amino acid sequences. Comparison with
herpesviral TKs reveals three conserved regions: glycine- and
arginine-rich ATP-binding motifs and a D/E-R-S/H motif at the putative
TK deoxynucleoside site. Greater homology, however, is seen upon
multiple alignment of dGK with mammalian deoxycytidine kinases,
yielding the consensus sequence -D/E-R-S-I/V-Y-x-D-. dGK also shares a
sequence (-Y-D-P-T-I/L-E-D-S/Y-Y-) required for GTP hydrolysis by
p21
.
)few appear to have any of the other
three deoxynucleoside kinases. Deoxycytidine kinase (dCK) has been
detected in extracts of Bacillus megaterium KM(3) , Bacillus subtilis(4) , and a number of species of the
class Mollicutes(5) , but none of the latter exhibited
ATP-dependent deoxyadenosine kinase (dAK), and only one (Spiroplasma ciri) contained deoxyguanosine kinase (dGK).
Extracts of pneumococci phosphorylate dAdo and dGuo, but apparently not
dCyd(6) . The only other bacterium clearly shown to contain all
four deoxynucleoside kinases, so far as we are aware, is B.
subtilis(7) in which dCyd and dAdo compete for a broadly
specific common active site. A spontaneous B. subtilis mutant
lacking dCK and dAK activities continues to grow, however.
)whereas the other three
deoxynucleoside kinase activities are uniquely organized into
heterodimeric proteins containing subunits highly specific for the
phosphorylation of dCK/dAK or dGK/dAK, respectively. The subunits of
each pair have very recently been separated, identified functionally,
and partially sequenced(9) ; the N-terminal sequences of all of
these subunits appear to be very similar but not identical.
Materials
The Photogene nucleic acid detection
kit and biotin-7-dATP were purchased from Life Technologies, Inc. The
Sequenase kit with modified T7 DNA polymerase was from U. S.
Biochemical Corp. [
-
S]dATP was purchased
from Amersham Corp. Restriction endonucleases and DNA modifying enzymes
were obtained from Boehringer Mannheim or Life Technologies, Inc. Taq polymerase was purchased from Amersham, Promega, or
Perkin-Elmer. Deoxynucleotides were from Perkin-Elmer. Plasmid DNA
purification cartridges were purchased from Qiagen. Phagemid
pBluescript(+)KS was from Stratagene. NENSORB-PREP oligonucleotide
purification cartridges were purchased from DuPont. Oligonucleotide
primers were synthesized in the Biochemical Instrument Center of The
Ohio State University.Bacterial Strains and Plasmids
L. acidophilus R26 (ATCC 11506, recently designated Lactobacillus sp. johnsonii) was used as the source of genomic DNA for the
recombinant library. The bacteria were cultured as described previously (14, 15) . The phagemid pBluescript(+)KS was used
for library construction and subcloning. Libraries were constructed
from purified size-selected restriction fragments of genomic DNA.Construction of a Biotinylated Cloning Probe by PCR and
Colony Hybridization
A DNA probe representing the coding
sequences of the N-terminal 26 amino acids was amplified by PCR with
thermostable Taq polymerase(16) . Genomic DNA (1
µg) was subjected to PCR for 30 cycles in a total volume of 50
µl containing 3 and 5 µM degenerate sense and
antisense primers, respectively, and 200 µM each of dNTPs
(dATP, dCTP, dGTP, and dTTP), 1 
buffer (50 mM KCl, 10
mM Tris-HCl at pH 8.4, and 2.5 mM MgCl),
and 1.25 units of Taq polymerase. The DNA samples were
denatured at 94 °C for 1 min, annealed at 37 °C for the first
five cycles and at 50 °C for the subsequent 25 cycles for 1 min,
and extended at 72 °C for 30 s. The desired 93-bp product was
extracted from agarose electrophoresis gel and reamplified by PCR under
the conditions described above, except for annealing at 50 °C
throughout all 30 cycles. Finally, using the reamplified PCR 93-bp
product (1 ng) as a template, PCR was repeated using 150 µM dATP and 50 µM biotin-7-dATP, for 20-30 cycles
of amplification.Genomic Library Screening by PCR
Colonies grown to
2-3-mm diameter on LB plates containing ampicillin were
resuspended in 50 µl of TTE buffer (1% Triton X-100, 20 mM Tris-HCl, pH 8.5, and 2 mM EDTA). Pooled suspensions of
10-20 colonies were screened by the PCR reaction. After boiling
for 5 min to lyse the cells and centrifuging, 5 µl of the
supernatant fraction were used directly to provide templates for PCR,
amplifying for 35 cycles (see ``Results'' for primers used).
In this manner, about 200 colonies could easily be screened in 10
reactions. Each colony of the 20-clone pool giving a positive PCR
reaction was screened individually by PCR and a clone selected.DNA Sequencing
Plasmid DNA templates used for
double-stranded sequencing were purified on a Qiagen tip-20 column.
Restriction fragments of the positive clones were subcloned into
pBluescript, and DNA sequences were determined using the
dideoxynucleotide chain termination method of Sanger et al.(18) , labeling the fragments with
[
P]dATP or
dATP--[S]thiophosphate.
of about 1.0,
harvested by centrifugation, resuspended in extraction buffer (0.1 M Tris-HCl, pH 8.0, 3 mM EDTA, 20% glycerol)
containing 2 mM phenylmethylsulfonyl fluoride, and opened by
sonication, on ice. Crude extracts were further purified by
streptomycin fractionation, ammonium sulfate fractionation, and
dATP-Sepharose affinity chromatography(15, 22) .
Construction of Cloning Probe by PCR
At the
outset of this work, the only amino acid sequence information available
on the Lactobacillus deoxynucleoside kinases was that of the
N-terminal 28 residues of the the combined dCK/dAK subunits (13) from enzyme purified to homogeneity, in very small
quantities, by dCTP-Sepharose affinity chromatography. Several factors
prevented the successful application of mixed oligonucleotide probes,
including highly degenerate codons, a conserved ATP-binding site near
the N terminus, and finding two different amino acids at several
residue positions in the sequence. By employing the PCR, a cloning
probe with most of the sequence degeneracy eliminated (and therefore
one with a higher degree of specificity) has been constructed on the
basis of that sequence. The sense primer for the PCR reaction,
5`-tgctctagATGATXGTN(CT)TN(AT)(GC)NGG-3`
(a 17-mer, degeneracy of 1536), comprised the coding sequence of amino
acids 1-6 plus an XbaI restriction site added to the 5`
end (N denotes T+C+A+G, X denotes
T+C+A). The antisense primer
5`-cggaattcT(TG)NGTNCCNA(GA)(GA)TA-(TC)TT-3`
(a 17-mer, degeneracy of 1024), contained the complementary coding
sequence of amino acids 26 to 21, with an EcoRI site added to
its 5` end. Thus, the putative template sequence in the genomic DNA
should be 77 bp, and the amplified PCR product, including the added
restriction sites, should be 93 bp. Nucleotide sequence analysis of the
PCR product showed that the sequence between the primers correlated
precisely with the amino acid sequences of dCK/dAK and was
nondegenerate. Within the primer regions, however, the sequence was
occasionally unclear, indicating utilization of multiple
oligonucleotides from the mixture by the PCR reaction. More recent
results, in which N-terminal peptide sequences of the separate subunits
have been determined and functional identities assigned, make it clear
that this probe could hybridize with the genes of any of the
four subunits of dCK/dAK or dGK/dAK since their N-terminal amino acid
sequences are nearly identical(9) . The 93-bp PCR product was
then labeled nonisotopically with biotin by incorporating biotin-dATP
during PCR amplification, as described under ``Experimental
Procedures.''Identification of Clones Containing Tandem
Genes
The biotinylated 93-mer PCR product was used as a probe in
the identification and isolation of one or more of the Lactobacillus deoxynucleoside genes. A 3.4-kb XbaI-digested genomic fragment hybridized strongly to the
probe. Colony hybridization revealed two positive clones out of more
than a thousand recombinants from an XbaI library, which was
constructed in the pBluescript vector in E. coli XL1-Blue
cells. Both clones were found to contain a 3.4-kb XbaI insert
that hybridized to the probe, and the nucleotide sequences revealed
apparently tandem genes, but with one-third of the 3` end of the
putative second gene missing from the XbaI fragment. No
activity could be detected toward any deoxynucleoside substrate, so we
could not determine which enzymes were represented by these partial
clones.Nucleotide Sequence Analysis
The coding and the
complementary strands of the entire 2.5-kb KpnI insert have
been sequenced. Sequencing strategies involved subcloning and
``gene-walking,'' using additional sequence-specific primers.
The nucleotide sequence revealed two open reading frames (ORF). The
nucleotide and derived amino acid sequences of the two genes are shown
in Fig. 1. The first ORF is 648 bp in length, with the 215
codons encoding a peptide with a predicted size of 25 kDa. The
N-terminal sequence encoded in this gene exactly matches the known
amino acid sequence of residues 2-28 of wild-type
dAK(9) . This gene is terminated by a single translation stop
codon (TAG) and is separated from the next ORF by a 21-bp spacer. The
next 675-bp ORF encodes a peptide of 224 amino acid residues with a
predicted molecular mass of 26 kDa. The N-terminal sequence inferred
from the DNA of this gene is identical to that determined for the
purified peptide of dGK, except for the initial methionine, but it also
matches the sequence of the dCK subunit residues
2-28(9, 13) . However, from a determination of
the activities expressed (see below) we find that we have cloned the
genes for the dGK/dAK pair.
Expression in E. coli
Lysates of E. coli XL1-Blue cells containing the pBluescript vector with or without
insert (control) were assayed for all three deoxynucleoside kinase
activities: dCK, dGK, and dAK. Enzyme activities of dGK and dAK, but
not dCK, were observed in the crude extract of the KpnI clone.
Addition of isopropyl-1-thio-
-D-galactopyranoside to
activate the lac promoter of the vector did not further
increase the kinase activities, indicating that the dAK/dGK was
expressed by means of its endogenous promoter. Specific activities for
both dAdo phosphorylation and dGuo phosphorylation in the crude extract
of the KpnI clone were more than 10-fold higher than in
protein extracted from wild-type Lactobacillus.
Sequence Alignment of dAdo Kinase and dGuo
Kinase
Sequences of the dAK and dGK genes cloned from Lactobacillus were analyzed by the computer software DNAStar,
revealing a 65% identity overall in the two DNA sequences. Furthermore,
alignment of amino acid sequences inferred from the dAK and dGK genes
reveals 61% identity between the two subunits in the 215 amino acids
which overlap (Fig. 3). Regions that are highly conserved in
other groups of proteins are marked in boldface type. Three
regions in each of the two peptide sequences seem to be highly
conserved in other phosphotransferase enzymes: (i) glycine-rich
sequences, characteristic of ATP-binding sites(13) , located
very near the N terminus of each peptide; (ii) sequences containing the
Asp-Arg-Ser motif, 78-80 amino acid residues from the N termini,
and which by analogy with many viral thymidine kinases (27) might be related to the nucleoside binding sites; and
(iii) arginine-rich regions, located about two-thirds of the way along
the length of each peptide, thought to participate in the binding of
ATP phosphate groups.
accession no. L33894) dCKs (which are nearly
identical to one another). These results are shown in Fig. 4.
Groups of residues in common are found over the entire sequence,
especially for dGK, and fairly close homology with the glycine- and
arginine-rich regions is seen, as with the viral TKs. Moreover, the DRS
motifs of the Lactobacillus dGK and dAK both align with a
considerably longer dCK sequence (beginning with a glutamate rather
than an aspartate, yielding the consensus sequence
D/E-R-S-I/V-Y-X-D). These similarities suggest that
the Lactobacillus dAK and dGK, in particular, bear a closer
relationship with mammalian deoxycytidine kinases than with various
thymidine kinases for which the homology lies mainly in the ATP-binding
motifs and the very limited DRS/H triad. Evidence for the functional
significance of the DRS motif in Lactobacillus dGK/dAK is the
subject of a companion study(30) .
and other Ras-like
proteins(31, 32) , as seen in Fig. 5. There is
nearly a perfect match between this dGK motif and that of the Ras
consensus, whereas only 4 of 9 dAK residues are identical with those of
the Ras motif. Perhaps the most significant observation to be made now
is the fact that the dGK sequence includes the threonine, which, in the
p21 sequence, is critical for GTP hydrolysis. While the
closely related structure, dGTP, is bound by dGK, there is no evidence
that it is hydrolyzed there. Thus, it would appear that this sequence
has been adapted to a somewhat different function in the Lactobacillus dGK.
26
KDa)(9, 22) . N-terminal amino acid sequences of each
subunit of dGK/dAK, as well as of dCK/dAK, have also been determined.
The technique of differential tryptic digestion of one subunit (with
the other subunit under the protection of its respective triphosphate
end product) has pinpointed the dGuo phosphorylation site to one
subunit, whereas the dAdo phosphorylation site is on the
other(9) .26 kDa upon
SDS-polyacrylamide gel electrophoresis (Laemmli system, pH 9.5), and it
is resolved into nonidentical subunits of 25 and 26 kDa by the
multiphasic zone electrophoresis 3328.IV buffer system at pH 6.8. 
)The dAK subunit has but one tryptophan, while dGK has two. has yielded both some predictable comparisons and
some surprising new relationships. Multiple alignment of the dAK and
dGK amino acid sequences with those of the thymidine kinases (not
shown) has revealed two sequences common to most purine
nucleotide-binding proteins (sites i and iii highlighted in Fig. 3), plus the very short DRS/H motif (site ii), which appears to be required for TK activity and
maps close to residues which, when mutagenized, affect the K
for thymidine in herpesviral thymidine
kinases(36, 37, 38) . However, except for
these conserved motifs, the bacterial dGK/dAK bears little homology
with the thymidine kinases from various sources. The highly conserved
arginine-rich region with the consensus sequence,
R-X-X-X-R-X-R, and found at amino
acid residue positions 140-146 (site iii) of dGK and
dAK, is a common element of all the deoxynucleoside kinases sequenced
so far and is also a conserved site in adenylate kinases (27, 39) The corresponding third Arg residue (Arg-118)
of chicken adenylate kinase has been shown to be essential for activity
in chicken adenylate kinase by the site-directed mutagenesis and NMR
studies of Yan et al.(40) . Replacement of Arg-138 by
methionine, or even lysine, while causing no apparent structural
perturbation compared with wild-type enzyme, resulted in inactivation
of the enzyme. Kinetic and structural results suggested that Arg-138
stabilizes the transition state, and to a small extent, the ground
state ternary complex, thus implicating this Arg as a candidate
involved in transferring the phosphoryl group. Based on refined
crystallographic data, Dreusicke et al.(41) also find
that the arginines in this region of porcine adenylate kinase bind
substrate phosphoryl groups. The arginines in this conserved site of Lactobacillus dGK/dAK may therefore have a similar function. for the hydrolysis of GTP, and that hydrolysis is
accompanied by a large conformational change affecting the entire loop.
In the realm of pure speculation, it is tempting to postulate that the
large one-directional effects on conformation and activity induced in
the dAK subunit upon binding dGTP or dGuo might be mediated through
contact with just such a site.
)))
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
ABSTRACT
This article has been cited by other articles:
|
|
 |
|
  M. Johansson, A. R. Van Rompay, B. Degreve, J. Balzarini, and A. Karlsson Cloning and Characterization of the Multisubstrate Deoxyribonucleoside Kinase of Drosophila melanogaster J. Biol. Chem., August 20, 1999; 274(34): 23814 - 23819. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Guo and D. H. Ives Functional Assignment by Chimera Construction of the Domain Affecting Heterotropic Activation of Deoxyadenosine Kinase from Lactobacillus acidophilus R-26 J. Biol. Chem., October 9, 1998; 273(41): 26624 - 26630. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Wang, J. Neuhard, and S. Eriksson An Escherichia coli System Expressing Human Deoxyribonucleoside Salvage Enzymes for Evaluation of Potential Antiproliferative Nucleoside Analogs Antimicrob. Agents Chemother., October 1, 1998; 42(10): 2620 - 2625. [Abstract] [Full Text]
|
|
|
|
|
|
M. Johansson and A. Karlsson Cloning of the cDNA and Chromosome Localization of the Gene for Human Thymidine Kinase 2 J. Biol. Chem., March 28, 1997; 272(13): 8454 - 8458. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Guo, N. Ma, and D. H. Ives cis-Active Ras G2-like Sequence Implicated in the Heterotropic Activation of the Deoxyadenosine Kinase of Lactobacillus acidophilus R-26 J. Biol. Chem., March 14, 1997; 272(11): 6890 - 6897. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Ma, S. Ikeda, S. Guo, A. Fieno, I. Park, S. Grimme, T. Ikeda, and D. H. Ives Deoxycytidine kinase and deoxyguanosine kinase of Lactobacillus acidophilus R-26 are colinear products of a single gene PNAS, December 10, 1996; 93(25): 14385 - 14390. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. S. Hong, G. T. Ma, and D. H. Ives Directed Mutagenesis of Deoxyguanosine Site at Arginine 79 Up-regulates Turnover on Deoxyadenosine Kinase Subunit of Heterodimeric Enzyme from Lactobacillus acidophilus R26 J. Biol. Chem., March 24, 1995; 270(12): 6602 - 6606. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |