ABSTRACT
The binding of transposase (Tnp) to the specific Tn
5
end sequences is the first dedicated reaction during transposition. In this
study
,
comparative DNA-binding analyses were performed using purified full-length Tnp and a C-terminal deletion variant (
[Delta]
369) that lacks the putative dimerization domain. The shape of the binding curve
of full-length Tnp is sigmoidal in contrast to the hyperbolic-shaped binding curve of
[Delta]
369. This observation is consistent with previous observations as well as a rate
of binding study presented here
,
which suggest that the full-length Tnp-end interaction
,
unlike that of the truncated protein
,
is a complex time-dependent reaction possibly involving a subunit exchange. Circular
permutation assay results indicate that both proteins are capable of distorting
the Tn
5
end sequences upon binding. Molecular weight determinations based on the
migratory patterns of complexed DNA in polyacrylamide gels has shown that
[Delta]
369 specifically binds the Tn
5
end sequences as a monomer while full-length Tnp in complex represents a heterodimer.
Tn
5
is a composite transposon composed of two insertion sequences: IS
50
R and IS
50
L that flank antibiotic resistance genes. The functional Tn
5
transposase (Tnp) is expressed from IS
50
R (Fig.
1
). The inhibitor protein (Inh), translated from the same reading frame, lacks
the N-terminal 55 amino acids. Tnp is a
cis
-active transposase that specifically recognizes two unique, 19 bp end
sequences (OE and IE) positioned at the termini of each IS
50
. With the exception of position 4, the first 9 bp of each site are identical.
The non-identical sections of each site contain binding sites for host proteins
(for a review see ref.
1
).
The critical initial step of Tn5 transposition requires the specific recognition
of end sequences by the Tnp protein. Two distinct Tnp-OE complexes have been observed in gel retardation assays: Complex I and
the faster migrating Complex II (
2
). Depending on the Tnp protein preparations used, wt Tnp or Tnp MA56 (which
eliminates production of the Inh protein), the proteins present in Complex I
include either Tnp, Inh and a naturally-occurring N-terminal deletion product, Tnp [alpha], or Tnp and Tnp [alpha], respectively (Fig.
1
). Since neither Inh nor Tnp [alpha] can themselves form specific nucleoprotein complexes with the OE, their
presence in Complex I indicates a protein-protein interaction with full-length Tnp. The migratory pattern of Complex II, shown to represent
binding of C-terminal deletion products, Tnp [gamma] and Tnp [delta] (Fig.
1
), is indicative of a change in the oligomeric states of these proteins when
complexed with the OE (
2
). The protein composition of Complex I as well as a number of other
observations have suggested that the full-length Tnp-OE interaction is a complicated reaction. Addition of Inh protein
stimulates the binding activity of Tnp to the OE presumably through a protein-protein interaction (
3
). In addition, a prolonged incubation time increases the amount of Tnp-OE complex formed (
2
). This may be related to a disaggregation and subunit exchange between full-length Tnp and Inh resulting in hetero-oligomer formation.
As has been demonstrated for a number of transposases from other systems (
4
-
6
), Tnp also distorts the OE DNA upon binding as determined by a circular
permutation assay (
7
). This distortion is the result of bending as determined by phasing analysis
(York and Reznikoff, in preparation). The apparent bending angle is >100o and centers near the first to third nucleotide of the 19 bp OE fragment (
7
).
Preliminary domain mapping of the Tn
5
Tnp protein was accomplished by deletion analysis. Various restriction
fragments of the Tnp gene were used in an
in vitro
transcription/translation experiment to generate a family of C-terminal variants of Tnp protein (
8
). DNA-binding analysis with these Tnp products have suggested that deletion of
the C-terminal 107 amino acids restricts Tnp oligomerization. This Tnp variant, [Delta]368, also demonstrated an apparent increased binding affinity. This
result implies that the C-terminus of the protein may partially block the DNA-binding domain of the protein, therefore reducing the overall DNA
binding affinity for the OE.
The goal of this paper is to further characterize the truncated variant of
transposase
in vitro
. A T7 expression vector was made to facilitate the overproduction of a
transposase variant lacking 107 amino acids from the C-terminus. The protein was purified using a modified procedure previously
reported. Purified preparations of this protein, designated [Delta]369, and full-length Tnp were tested in gel retardation assays for binding
affinity and ability to bend the OE. Molecular weight determinations based on
the migration patterns of complexed DNA in polyacrylamide gels using both the
truncated and full-length protein were also performed which indicated that Complex I
represents a heterodimer of Tnp bound to a single OE-DNA fragment while [Delta]369 binds specifically as a monomer.
Escherichia coli
strains DH5[alpha] and BL21(DE3) pLysS were used for plasmid isolation and transposase overexpression, respectively. Plasmids pRZ7074 MA56, pRZ7067 (
2
), pRZ4826 (
8
) and pRZ9012 (
7
) were described previously. pRZ9000 (this study) is described below.
Plasmid pRZ9000 used for overexpression of the truncated Tnp, was constructed
from pRZ7074 (
2
) and pRZ4826 (
8
). pRZ7074 (
2
), is a pET21d derivative which contains the entire Tnp gene (with MA56 mutation
that eliminates Inh production) under control of the IPTG-inducible T7 promoter. Digestion with
Hin
dIII results in two fragments: a 6448 bp fragment containing the T7 promoter
region and ampicillin resistance gene and the amino-half of the Tnp gene (encoding the first 368 amino acids) and a 1900 bp
fragment containing the carboxy-half of the Tnp gene and the rest of the vector. These fragments were
treated with Mung Bean nuclease (New England Biolabs) according to
manufacturer's instructions and the large fragment was isolated and ligated to
a 593 bp
Sma
I-
Nru
I DNA fragment, isolated from plasmid pRZ4826 (
8
), which includes stop codons in all three reading frames at the 5' end. The resulting construct, pRZ9000, was confirmed by sequence analysis and used for the overexpression of the truncated form of Tnp [368 amino acids from Tnp and an additional
codon (gly) from the stop cassette].
An overnight culture of BL21 (DE3) pLysS containing pRZ9000 MA56 (eliminates
production of Inh protein) was used to inoculate 2 l of tryptone-phosphate broth (2% bacto-tryptone, 0.2% Na
2
HPO
4
, 0.1% KH
2
PO
4
, 0.8% NaCl and 2% yeast extract) (
9
). Cells were grown at 37oC to an OD
600
of 0.5. Protein overexpression was induced with IPTG at a final concentration
of 0.1 mM. After an additional 1.5 h incubation, cells were harvested and the
protein purified as previously described (
10
). The homogeneity of [Delta]369 was analyzed on a denaturing 10% SDS-PAGE gel followed by Coomassie staining. Full-length Tnp protein, overexpressed and purified from BL21
(DE3) pLysS cells containing pRZ7074 MA56, was a gift from Maggie Zhou. This
full-length Tnp preparation, as analyzed on a Coomassie-stained SDS-PAGE, was determined to be ~90% pure (Fig.
2
).
The binding affinity assay was performed using an [alpha]-
32
P-labeled 52 bp
Eco
RI-
Hin
dIII fragment from pRZ7074 MA56 (
2
). Reaction mixtures of 15 [mu]l [100 mM potassium glutamate, 25 mM Tris-acetate (pH 7.5), 10 mM magnesium acetate, 0.5 mM [beta]-mercaptoethanol, 400 mg/ml bovine serum albumin (BSA), 600
ng tRNA and 20% glycerol] containing 1 nM (0.5 ng) of labeled probe were
incubated for 30 min at 30oC with increasing amounts of purified full-length Tnp MA56 and [Delta]369 MA56 (0, 62.5, 125, 625 and 1250 nM). 3.5 [mu]l of 20% glycerol/0.1% bromophenol blue/0.1% xylene cyanol
FF was added to each sample and loaded onto an 8% non denaturing 0.5* TBE polyacrylamide (29:1) gel. The gel was run at 300 V, 4oC for 2 h, dried and exposed to film overnight. The percentage of
free and bound DNA was determined by a Molecular Dynamics Phosphorimager, ImageQuant software.
For the rate of binding studies, 625 nM amounts of full-length Tnp and [Delta]369 were incubated for various times (0, 15, 30, 60 and 120 min)
with 1 nM (0.5 ng) of the 52 bp OE probe in 15 [mu]l of binding buffer (see above). Samples were processed as described above.
Molecular weight determination was performed as described by Orchard and May (
11
). The 52 bp OE fragment was incubated with 500 nM of full-length Tnp and [Delta]369, as described above, and run on a series of 0.5* TBE polyacrylamide gels (5-10%). Five micrograms of protein standards:
lactalbumin (MW 14 200), carbonic anhydrase (MW 29 000), chicken egg albumin
(MW 45 000), and BSA, which runs as a monomer (MW 66 000) and a dimer (MW 132
000) were also run simultaneously on each set of gels. For each gel the
position of the bromophenol blue dye was marked. The half of each gel
containing the binding reactions was dried and exposed to film overnight while
the protein standard half of each gel was stained with Coomassie. The migration
distances of each complex as well as each protein standard was measured and
divided by the distance the bromophenol blue migrated in the same sample
(relative mobility =
R
f
). The calibration curves from the known protein standards as well as those of
the full-length Tnp-OE and [Delta]369-OE complexes were derived from a plot of 100[log(
R
f
* 100)] as a function of gel concentration (%). The Ferguson plot is
derived from the negative slopes of the calibration curves of each of the
protein standards as a function of the known molecular weights and used to
determine the molecular weight of Tnp protein complexes.
Gel retardation assays were performed with four circularly permutated 182 bp
fragments from pRZ9012 containing the 19 bp OE consensus sequence as described
previously (
7
). The apparent bend angle was computed by using the formula of Zhou
et al
. (
13
), a derivative of the Thompson and Landy bending equation (
14
). This equation is as follows;
{{roman {mu sub italic i}} over roman {mu sub italic j}} {roman =} {{{roman {[ 1
- 2 ( {chi sub italic i} . / L ) ( 1 - c o s alpha ) + 2 ( {chi sub italic i} {{. / L )} sup 2} ( 1 - c o s alpha {{) ]} sup {0 . 5}}}}} over {{roman {1 - 2 ( {chi sub italic j} .}} {roman /} {roman {L ) (
1 - cos alpha ) + 2 ( {chi sub italic j} {{. / L )} sup 2} ( 1 - c o s alpha {{) ]} sup {0 . 5}}}}}}
where [mu]
i
and [mu]
j
are the relative mobilities of the complexes with Tnp bound to a centrally
positioned OE and a peripherally positioned OE respectively, [chi]
i
/L and [chi]
j
/L are the fractional distances from the center of bending (+3) to the left end
of the bending probes of the same OEs, and [alpha] is the angle of the bend.
The binding properties of the Tnp variant lacking the C-terminal 107 amino acids suggested that the protein alters both DNA-binding activity and oligomerization. Since DNA-binding and Tnp oligomerization are essential steps in
transposition, this study focused on more detailed biochemical analyses of this
truncated Tnp variant. A T7 overexpression vector containing the portion of the
Tnp MA56 gene (which eliminates Inh production) encoding the first 368 amino
acids was constructed (pRZ9000) as described in Materials and Methods. This
truncated protein, designated [Delta]369, when overexpressed under conditions previously outlined for full-length Tnp (
10
), formed insoluble inclusion bodies (data not shown). Denaturation and
renaturation of these insoluble aggregates were not successful therefore
attempts were made to increase the solubility of the protein by modifying
overexpression conditions. Using a procedure outlined by Moore
et al
. (
9
), [Delta]369 was overexpressed in BL21 (DE3) pLysS cells grown in a tryptone-phosphate broth as described in Materials and Methods. The soluble
fraction of the [Delta]369 protein, ~10-25% of the total [Delta]369 protein (Fig.
2
), was purified by a previously reported protocol (
10
). The homogeneity of the peak fraction was determined by a denaturing SDS-PAGE gel to be ~85% (Fig.
2
).
Purified preparations of full-length Tnp MA56 (lacking Inh protein) and [Delta]369 MA56 were tested for DNA-binding activity by a gel retardation assay. Increasing
amounts of full-length Tnp and [Delta]369 protein preparations were incubated with a 52 bp fragment
containing the 19 bp OE consensus sequence as described in Materials and
Methods. As shown in Figure
3
A, full-length protein forms the characteristic Complex I and the faster migrating
Complex II (
2
), while [Delta]369 forms a single faster migrating complex. This result mimics that
reported for the
in vitro
transcribed/translated 368 amino acid residue long C-terminal deletion variant (
8
).
The DNA-binding activity of each protein was compared by generating a binding
curve from a plot of the percent disappearance of OE DNA as a function of
protein concentration (Fig.
3
B). The relative binding affinities,
K
observed
(
K
obs
), of each protein preparation was determined from the protein concentration at
which 50% of the OE DNA was bound. Surprisingly, the
K
obs
of the [Delta]369 protein preparation is ~2-fold lower than that of full-length Tnp protein. This result was unexpected based on
the previous binding analysis of
in vitro
transcription/translation products in which a nearly identical truncated
protein demonstrated an apparent higher binding activity than that of full-length Tnp (
8
). This discrepancy will be addressed in the Discussion.
Previous work had suggested that formation of Complex I increases with prolonged
incubation (
2
). This observation is consistent with a subunit exchange between full-length Tnp aggregates and Tnp [alpha] or Inh. We decided to further explore this possibility by
incubating full-length Tnp MA56 and [Delta]369 with OE DNA. Binding reactions were carried out as outlined in
Materials and Methods, however, incubation times were varied (0, 15, 30, 60 and
120 min). As shown in the gel retardation analysis in Figure
4
, the formation of full-length Tnp-OE complexes increases with longer incubation times. In contrast,
>50% of the [Delta]369-OE complexes form almost instantaneously with the full extent of
binding occurring within 15 min. It is of interest that the quantity of
specific [Delta]369-OE complex formation decreases with longer incubation times which
may be a reflection of the instability of [Delta]369-OE complex based in part on the inherent ability of the protein to
aggregate.
The oligomeric state of full-length Tnp MA56 (lacking Inh protein) and [Delta]369 MA56 in complex with the OE was determined using a modified
Ferguson analysis reported by Orchard and May (
11
,
12
). This method allows the molecular weight of protein-DNA complexes to be determined indirectly by running them on a series of non-denaturing polyacrylamide gels of differing acrylamide concentrations alongside a number of protein standards of known
molecular weights. Mobility is therefore related to the sieving effects of the
gel and consequently to the size and shape of the protein in complex. The
molecular weight of an unknown protein in a DNA complex can be determined
graphically from mobility differences.
The 52 bp fragment containing the 19 bp OE consensus sequence was incubated with
a single concentration of full-length Tnp and [Delta]369 under the conditions described in Materials and Methods.
Reactions were run on a series of polyacrylamide gels (5-10%) alongside a set of protein standards and processed as described
under Materials and Methods. The relative mobility (
R
f
) of Complex I, Complex II and [Delta]369-OE complex as well as each protein standard was determined for
each gel concentration. A calibration curve was derived for each of the protein
standards and each of the Tnp-OE complexes (Fig.
5
A). The gradient of the calibration curve of the protein standards was plotted
as a function of their known molecular weights on a log scale (Ferguson plot).
From this linear plot (Fig.
5
B) the approximate molecular weight of the protein-DNA complexes was determined and after subtracting the contribution of
the OE DNA (MW 36 075), the molecular weight of the protein in complex was
approximated. The protein component of Complex I (MW 110 000) represents a
homodimer of full-length Tnp (MW 53 311) and/or a heterodimer of full-length and Tnp [alpha] (MW ~49 000/monomer). The protein component of Complex II (MW
44 000) most likely represents a monomeric form of Tnp [delta] and/or [gamma]. The molecular weight of [Delta]369 (MW 41 473 based on amino acid content) in complex is ~41 000 which coincides with a monomeric form of the
protein.
Recently it was shown that full-length Tnp bends the OE upon binding at an apparent bend angle >100o, centered at position 3 (
7
). Comparison analysis based on a circular permutation assay was performed using
[Delta]369 protein as well as full-length Tnp. Four circularly permuted 182 bp fragments containing
the 19 bp OE segment at various positions (Fig.
6
A) were used in a gel retardation analysis with full-length and [Delta]369 protein. As seen in Figure
6
B, mobility of complexes is dependent on the position of the OE sequence in the
fragment, an indication that both proteins bend the OE. The
R
f
of Complex I (full-length Tnp MA56) and the [Delta]369-OE complex was measured and plotted as a function of the
position of the center of each of the bending probes (Fig.
6
A). These curves (Fig.
6
C) allowed us to map the center of the full-length Tnp-induced and [Delta]369-induced bend, located at the lowest point of the curve.
For both these proteins, the lowest point is at position 135 which maps
approximately to the third nucleotide in the 19 bp OE sequence.
Tn
5
transposase is a multifunctional protein that is an essential component in all
the steps involved in transposition: site-specific binding to end sequences, synaptic complex formation through
oligomerization, double-stranded cleavage of the DNA immediately adjacent to the Tn
5
ends and subsequent cleavage of the target DNA, and strand exchange of the free
3' ends to the 5' ends of the target. Preliminary domain mapping has suggested that
the N-terminus contains the determinants for specific end recognition while the
oligomerization domain lies in the C-terminus of the protein (
8
; Zhou and Reznikoff, in preparation). Amino acid homologies with other
transposase and integrase proteins (
16
) suggest that the catalytic domain of Tnp is located near the central portion
of the protein.
Previous binding analyses of a family of C-terminal deletion variants of Tnp has shown that deletion of the last 107
amino acids changes the oligomeric state of Tnp in complex with the OE. In
addition, this deletion variant of Tnp demonstrated a higher apparent binding affinity for the OE suggesting that the C-terminus of the protein partially inhibits Tnp binding (
8
). This study focuses on further
in vitro
characterization of this protein variant. This truncated form of Tnp, [Delta]369, was purified and analyzed
in vitro
in a comparison study with full-length Tnp protein. DNA-binding ability of each protein, tested in a gel retardation assay
(Fig.
3
), showed that there is only a 2-fold difference in the relative binding affinity of each protein
preparation. This result is in sharp contrast with that reported in the
deletion studies (
8
). However, this can be explained three different ways: (i) the protein
concentrations used in the previous study may fall in the early part of the
curve (Fig.
3
B) where it appears that [Delta]369 has a four to five times higher binding affinity for the OE in
comparison with full-length Tnp protein, (ii) since full-length Tnp binding is time dependent (Fig.
4
), the incubation time used in the previous study may not have been sufficient
to observe the full extent of binding, and (iii) the availability of N-terminal proteolytic products, which seem to have a direct effect on full-length Tnp-DNA complex formation, may be significantly lower in the Tnp
protein preparation produced in the eukaryotic
in vitro
transcription/translation system.
In addition to similar relative binding affinities, both proteins show the same
severe distortion of OE DNA upon binding as demonstrated by the circular
permutation assay. This distortion has recently been shown by a phasing
analysis to be a direct result of protein-induced bending (York and Reznikoff, in preparation). Bending of the OE by
either protein is ~116o with the bend center located at the third position in the 19 bp
sequence. This result is in agreement with that reported previously with full-length protein (
7
). An asymmetrical bend center relative to a protein binding site has also been
demonstrated for the TnsB protein of Tn
7
(
4
). The proximity of the Tnp-induced bend center near the first nucleotide in the OE strongly suggests
that additional nucleotide contacts are needed upstream for efficient binding
and bending by Tnp. Experiments to determine the minimal binding site needed
for optimal protein binding to the OE are currently in progress.
While both full-length Tnp and [Delta]369 exhibit strong similarities in a number of the
in vitro
assays presented here, two significant differences were observed. One
difference involves the shape of the binding curve (Fig.
3
B). The hyperbolic shape of the binding curve of [Delta]369 is generally representative of a simple one step binding reaction in
which a single preformed protein component binds to DNA. However, the sigmoidal
curve binding of full-length Tnp suggests a more complex reaction. This hypothesis is supported by a number of previous observations. Protein composition
studies of Complex I have shown the presence of full-length Tnp, Inh and Tnp [alpha], a naturally-occurring proteolytic product lacking ~25 amino acids from the N-terminus, when using wild-type Tnp preparations, or full-length Tnp and Tnp [alpha] when using Tnp MA56 (as in
this study). Full-length Tnp protein appears in an equimolar ratio with the other
polypeptides (
2
,
3
). Since Tnp [alpha] does not possess sequence-specific binding abilities (
3
), its presence in Complex I indicates that full-length Tnp forms complexes with Tnp [alpha] through a protein-protein interaction. Based on the molecular weight
determination presented in this study with Tnp MA56 (lacking Inh protein), this
association is in the form of heterodimers: full-length Tnp/Tnp [alpha]. The molecular weight determination does not rule out the
possibility that full-length Tnp homodimers also bind to OE DNA. Experiments are currently
underway to determine what different dimeric species of Tnp make up Complex I
and the relative amounts of each that are present. This information would give
us an indication of the relative affinity of each Tnp dimeric species for OE
DNA. The second important difference between full-length Tnp and [Delta]369 involves the time required for each protein to form complexes
with the OE. Full-length Tnp complex formation steadily increases over time. In contrast, [Delta]369-OE complex formation maximizes very rapidly. Again this result
suggests that a subunit exchange occurs between full-length Tnp aggregates and Tnp [alpha].
We believe that the formation of Tnp heterodimers through a subunit exchange is
the key to the complexity of full-length Tnp-OE binding reflected in part by the sigmoidal shape of the binding
curve and in part by the slow formation of full-length Tnp complexes. Equilibrium-binding studies with the
Eco
RI restriction endonuclease in the presence of DNA lacking
Eco
RI recognition sites displayed a sigmoidal-shaped curve. This sigmoidicity was shown to reflect the addition of
another enzyme subunit (
17
). A model involving Tnp subunit exchange is supported by a number of
observations. Previous experiments have shown that titration of reactions
containing constant amounts of Tnp MA56 protein preparations and OE DNA with
Inh protein results in a steady increase in Complex I formation. This is
presumably due to the increased availability of Inh to form heterodimers with
full-length Tnp. In addition, as shown in this study, longer incubation times
also increases Complex I formation (Fig.
4
). This implies that heterodimer formation and subsequent OE binding are dynamic
processes limited by the availability of Inh and/or Tnp [alpha]. This hypothesis also implies that the heterodimer is more active in
binding than homodimers of full-length Tnp or that homodimers of full-length Tnp are trapped in aggregates under the conditions used in
the assays. The possible role of heterodimers in Tn
5
transposition is discussed below.
Heterodimer formation between a number of eukaryotic transcriptional regulatory
proteins that share dimerization motifs have been shown to be involved in the
regulation of their activity by modifying their affinity for different DNA
targets (for a review see ref.
18
). In addition, heterodimer formation has also been shown to act as an
inhibition mechanism. Several heterodimer forms of the mammalian
transcriptional activator protein, CREB, act as repressors of cAMP-induced transcription while the monomeric form acts as a strong
constitutive activator (
19
). This model is not unlike that proposed for the regulation of Tn
5
transposition (
2
) as described below.
The mechanism of transposition is a highly regulated process. In Tn
5
, the Tnp protein has two opposing activities: (i)
in cis
the protein is able to catalyze transposition (
20
) and (ii) Tnp acts primarily as an inhibitor
in trans
. The Inh protein acts exclusively as an inhibitor of transposition (
10
). It is proposed that immediately after translation, Tnp exists briefly as the
active monomeric form which is able to bind an end and mediate synapse
formation through dimerization of a second end-bound monomer. This intermediate is competent for subsequent steps of
transposition. The Tnp protein soon after translation is subject to
dimerization and possibly proteolysis
in vivo
, both of which may contribute to inhibition of transposition. All the
proteolytic products of Tnp as well as the Inh protein have been found to be
inactive in promoting transposition
in vivo
(
2
,
3
). However, all may contribute to inhibition. This mechanism of inhibition can
occur by two not mutually exclusive steps. The active Tnp monomers form
inactive homodimers and/or heterodimers with Inh or a proteolytic product. This
dimer formation reduces the active population of Tnp. In addition, these
dimeric isoforms can specifically bind to end sequences blocking access of the
active monomeric form of Tnp.
The mechanism of Tnp heterodimer formation is currently under investigation.
Preliminary observations suggest that the subunit exchange indicated by the
sigmoidal binding curve as well as the rate of binding study occurs primarily
in the presence of OE DNA (York and Reznikoff, data not shown). This is not
unlike that seen with CAP protein (
21
). This line of investigation may give us a better insight into the mechanism of
transposition inhibition.
This work was supported by NIH grant GM50692 and a gift from the Promega
Corporation. D.Y. was supported by NCI training grant T32 CA09075. W.S.R. is
the Evelyn Mercer Professor of Biochemistry and Molecular Biology.
REFERENCES
Return