High-resolution physical and functional mapping of the template adjacent DNA binding site in catalytically active telomerase.

Abstract

Telomerase is a cellular reverse transcriptase, which utilizes an integral RNA template to extend single-stranded telomeric DNA. We used site-specific photocrosslinking to map interactions between DNA primers and the catalytic protein subunit (tTERT) of Tetrahymena thermophila telomerase in functional enzyme complexes. Our assays reveal contact of the single-stranded DNA adjacent to the primer-template hybrid and tTERT residue W187 at the periphery of the N-terminal domain. This contact was detected in complexes with three different registers of template in the active site, suggesting that it is maintained throughout synthesis of a complete telomeric repeat. Substitution of nearby residue Q168, but not W187, alters the K(m) for primer elongation, implying that it plays a role in the DNA recognition. These findings are the first to directly demonstrate the physical location of TERT-DNA contacts in catalytically active telomerase and to identify amino acid determinants of DNA binding affinity. Our data also suggest a movement of the TERT active site relative to the template-adjacent single-stranded DNA binding site within a cycle of repeat synthesis.

Full text

Page 1

High-resolution physical and functional mapping
of the template adjacent DNA binding site
in catalytically active telomerase
Erez Romi*, Nava Baran*, Marina Gantman*, Michael Shmoish†, Bosun Min‡, Kathleen Collins‡, and Haim Manor*§
Departments of *Biology and†Computer Science, Technion–Israel Institute of Technology, Haifa 32000, Israel; and‡Department of Molecular and Cell
Biology, University of California, Berkeley, CA 94720-3204
Communicated by E. Peter Geiduschek, University of California at San Diego, La Jolla, CA, April 5, 2007 (received for review October 20, 2006)
Telomerase is a cellular reverse transcriptase, which utilizes an
integral RNA template to extend single-stranded telomeric DNA.
We used site-specific photocrosslinking to map interactions be-
tween DNA primers and the catalytic protein subunit (tTERT) of
Tetrahymena thermophila telomerase in functional enzyme com-
plexes. Our assays reveal contact of the single-stranded DNA
adjacent to the primer-template hybrid and tTERT residue W187 at
the periphery of the N-terminal domain. This contact was detected
in complexes with three different registers of template in the
active site, suggesting that it is maintained throughout synthesis
of a complete telomeric repeat. Substitution of nearby residue
Q168, but not W187, alters the Kmfor primer elongation, implying
that it plays a role in the DNA recognition. These findings are the
first to directly demonstrate the physical location of TERT-DNA
contacts in catalytically active telomerase and to identify amino
acid determinants of DNA binding affinity. Our data also suggest
a movement of the TERT active site relative to the template-
adjacent single-stranded DNA binding site within a cycle of repeat
synthesis.
specific cleavage of proteins ? telomerase–primer interaction ?
UV crosslinking
T
copying a template within the integral RNA component of the
enzyme (1). Some telomerase enzymes can also use this internal
template to direct the synthesis of telomeres at nontelomeric
sites of chromosome fragmentation (2). In addition to the
telomerase RNA subunit (TER), the enzyme contains a catalytic
protein subunit, designated telomerase RT (TERT), and acces-
sory proteins (3, 4).
Telomerase was first discovered in extracts of the ciliate
Tetrahymena thermophila (5), and telomerase from this organism
remains an excellent model system for studies of enzyme struc-
tureandfunction.ItsRNAsubunit(tTER)of159ntcontainsthe
repeat-complementary sequence 3?-AACCCCAAC-5? and
othermotifsrequiredforribonucleoprotein(RNP)assemblyand
activity (1, 3). T. thermophila TERT (tTERT) consists of 1,117
amino acids, including a region between residues 518 and 881
that is conserved among RTs and designated as the RT domain
(6). The N-terminal half of TERT contains motifs conserved
among TERTs but not viral RTs. It constitutes two indepen-
dently folded domains: the TERT essential N-terminal domain
(TEN) and the TERT high-affinity TER binding domain
(TRBD). In tTERT, residues 1–195 can be considered to con-
stitute the TEN domain, whereas residues 196–528 comprise the
TRBD (7–9).
Telomerase specificity of interaction with single-stranded
DNA has been studied by monitoring the elongation of primers
of varying lengths, sequences and concentrations. Differences in
the primer concentration-dependence and repeat addition pro-
cessivity of product synthesis indirectly suggest that extensive
contacts to the enzyme are made by primer regions 5? of the
elomerase is a unique reverse transcriptase (RT) that ex-
tends the single-stranded 3? overhangs of telomeres by
template hybrid (2). More direct physical assays have also been
used to investigate enzyme–primer interactions. Our previous
interference footprinting studies indicated that functionally non-
redundant interactions of primer with enzyme occur primarily in
the six or seven 3?-terminal primer nucleotides (10). In addition,
atomic-resolution structure determined for residues 13–176 of
the tTERT TEN domain revealed a surface groove with features
suggestive of a channel for binding single-stranded DNA (11).
Mutagenesis of some channel residues (Q168A, F178A) strongly
reduced recombinant telomerase activity, and activity was elim-
inated by an adjacent substitution (W187A) in the C-terminal
‘‘tail’’ of TEN that adopted alternative structures. These same
substitutions each reduced the specificity of TERT crosslinking
to a radiolabeled, iodouracil-derivatized DNA primer (11).
Here we investigate the sites of tTERT interaction with DNA
by mapping covalent crosslinks induced by UV light. For func-
tional significance, we designed the assays to map tTERT
interaction site(s) for a nucleotide of the primer at the boundary
of the primer-template hybrid in catalytically active telomerase
complexes (see below). These interaction site(s) can be studied
most readily with recombinant core enzyme containing tTER
and tTERT, because the important template-adjacent DNA
contacts are not overshadowed by the additional DNA interac-
tions that may occur in telomerase holoenzyme complexes (12).
We characterized a site-specific DNA crosslink to tTERT tryp-
tophan 187 (W187). Another crosslink site was also detected
within the tTERT segment spanning residues 192–411 of the
TRBD. Primer extension activity assays revealed that W187 by
itself was not essential for recombinant enzyme activity, but
substitutions in the neighboring residue Q168 altered primer Km
for elongation. Our data are the first to directly map a specific
TERT–DNA interaction at a single amino acid resolution within
catalytically active telomerase RNP. Overall, the data provide
new information about the positioning of DNA and TERT
domains in functional telomerase complexes.
Results
Mapping of DNA Crosslinking Sites in Catalytically Active Telomerase.
We performed site-specific DNA crosslinking to TERT, using
the 6-nt primer 5?-G(IdU)TGGG-3? (IdU5), in which a thymi-
dine was substituted with 5-Iododeoxyuridine (IdU) at the (?5)
Author contributions: E.R., K.C., and H.M. designed research; E.R., N.B., M.G., and B.M.
performedresearch;E.R.,N.B.,M.S.,K.C.,andH.M.analyzeddata;andK.C.andH.M.wrote
the paper.
The authors declare no conflict of interest.
Abbreviations: RT, reverse transcriptase; TEN, TERT essential N-terminal domain; TER,
telomerase RNA subunit; tTER, Tetrahymena telomerase RNA subunit; TERT, telomerase
catalytic protein subunit; tTERT, Tetrahymena telomerase catalytic protein subunit; TRBD,
TERT high-affinity TER binding domain.
§To whom correspondence should be addressed. E-mail: manor@tx.technion.ac.il.
This article contains supporting information online at www.pnas.org/cgi/content/full/
0703157104/DC1.
© 2007 by The National Academy of Sciences of the USA
www.pnas.org?cgi?doi?10.1073?pnas.0703157104PNAS ?
May 22, 2007 ?
vol. 104 ?
no. 21 ?
8791–8796
BIOCHEMISTRY

Page 2

position relative to the 3?OH end. This position was previously
shown to be involved in primer interaction with telomerase (10,
13). A complex of reconstituted core enzyme (tTERT and
tTER) with primer was irradiated with a long-wavelength UV
light (?295 nm), thereby favoring crosslinks of the IdU with
aromatic amino acids in close proximity. Next,32P-dGTP was
added to the mixture and the crosslinked DNA was extended
with a single32P-dGMP by the enzyme molecule to which it was
bound (Fig. 1A). Thus, only the crosslinked primer molecules
that were properly aligned within a functional RNP were radio-
labeled. The reactions were analyzed for tTERT crosslinking
(Fig. 1B Upper) and for enzyme activity (Fig. 1B Lower). In
complete reactions with substituted primer and UV, a single
radiolabeled protein having the expected mobility of tTERT was
observed, and the expected 7-nt extension product was detected
as well (lane 1). Similar experiments performed with unsubsti-
tuted primer, or without UV treatment, gave barely detectable
signal of crosslinked tTERT, or no signal, despite efficient
elongation (lanes 2 and 4). A reaction carried out in the absence
of primer produced no detectable radiolabeled protein or DNA
(lane 3), as did reactions carried out with radiolabeled dCTP
instead of dGTP (lane 5) or with tTER or tTERT missing from
the enzyme reconstitution (lanes 6–7). Taken together, these
results demonstrate that the observed crosslinking product
reflects site-specific interaction of the primer and tTERT in a
catalytically active complex.
The yield of crosslinked products obtained in our assays (?1
fmol) was too low to allow definition of the crosslinking sites by
using mass spectrometry. Therefore, we used a mapping method
involving partial protein degradation that was previously used in
studies of RNA polymerase elongation complexes (14, 15). After
primer crosslinking and radiolabeling by nucleotide addition,
tTERT molecules were immunopurified and digested with cy-
anogen bromide (CNBr). This reagent cleaves polypeptides on
the carboxyl side of methionine residues, of which there are 18
in tTERT (Fig. 1C). Partial digestion was performed such that
most tTERT molecules were either cleaved once or not cleaved
at all (‘‘single-hit’’ cleavage). Each of the tTERT molecules
cleaved at a single site was expected to yield two fragments, of
which only the fragment containing the crosslink would be
radiolabeled. Assuming that single-hit digestion occurred with
equal probability at all methionines of tTERT, the ensemble of
digested molecules should encompass 18 pairs of radiolabeled
and unlabeled fragments. Crosslinking at each segment of
tTERT delineated by sequential methionines should generate a
unique pattern of32P-labeled fragments, which was simulated by
computer as shown in Fig. 1D.
Partial digestion was performed over a time course and the
ratio of uncleaved and cleaved tTERT molecules was deter-
mined to most closely approximate conditions of single-hit
digestion (Fig. 1E, lanes 2–3). The resolution of the SDS/PAGE
gels used here was optimal in the range of fragments of 20–55
kDa. Smaller fragments were masked by the large amounts of
radioactivity from uncrosslinked product DNA and unincorpo-
rated dGTP (data not shown).
Comparison of experimentally observed fragments in Fig. 1E
to the simulated fragment patterns in Fig. 1D indicated com-
patibility only with a crosslinking site in the segments of amino
acids 2–13 or 14–194. The corresponding region in the yeast
TERT Est2p has been also proposed to interact with DNA
primers (16). The presence of a weakly labeled fragment in lane
1(undigestedtTERTcontrol)thatcomigratedwiththe26.0-kDa
fragment in lanes 2–4 prevented the unambiguous attribution of
the 26.0-kDa fragment to a digestion product of tTERT. If
digestion did not produce a 26.0-kDa fragment, then the
crosslinking site could map either in the segment 2–194, or in the
segment 195–235.
Subsequently, we carried out another experiment in which
Fig. 1.
ofthecrosslinkingexperiment.xdesignatestheIdUsubstitution.Xdesignates
the crosslinked IdU. g is the newly added [32P]-dGMP. Nucleotide numbers at
the boundaries of the RNA template region are indicated. (B) (Upper) SDS/
PAGEanalysisofcrosslinkedtTERT.Sizemarkersaredepictedontherightwith
masses in kilodaltons. (Lower) Analysis of the noncrosslinked DNA products.
Samples from the crosslinking reactions were withdrawn, and noncrosslinked
extended DNA primers (?99% of the products) were analyzed by urea-
polyacrylamide gel electrophoresis. (C) CNBr cleavage map of tTERT. CNBr
cleavage occurs at methionine residues, whose positions are indicated. (D)
Simulation of the SDS/PAGE patterns expected in a CNBr single-hit digestion
ofcrosslinkedtTERT.The18lanespresentthegelpatternsexpectedforthe18
possible locations of the crosslinking site. Size markers are depicted along the
y axis, whereas the cleavage sites are depicted along the x axis. Each lane is
located between two successive cleavage sites and reflects the gel pattern
expectedifthecrosslinkwouldoccurbetweenthesetwosites.Thesimulation
refers to fragments larger than 20 kDa, which were resolvable by SDS/PAGE.
The arrowheads point at fragments that were useful for mapping of the
crosslinking site. The simulation tool was implemented as the R-function (25)
andisavailableuponrequestfromM.S.(E)SDS/PAGEoffragmentsgenerated
by partial digestion of crosslinked tTERT with CNBr. The arrowheads indicate
the fragments that provided the map position of the crosslinking site. The
fragments are marked by their masses in kilodaltons. (F) SDS/PAGE of frag-
ments generated by extensive CNBr digestion of WT or M194L crosslinked
tTERT. The arrows indicate fully and partially digested fragments. The frag-
ments are marked by their map positions in tTERT.
Site-specific crosslinking of telomerase and DNA primer. (A) Scheme
8792 ?
www.pnas.org?cgi?doi?10.1073?pnas.0703157104 Romi et al.

Page 3

tTERT was extensively digested with CNBr (Fig. 1F). Under
these conditions, the majority of the tTERT molecules were
digested at all of the cleavage sites. This digestion gave rise to a
major radiolabeled fragment of ?23.5 kDa, which is the ex-
pected mass of the 14–194 peptide crosslinked to the elongated
primer. Also produced was a minor 28.5-kDa fragment, which is
the expected mass of the 14–235 peptide crosslinked to the
elongated primer and likely results from incomplete digestion at
M194. To further validate these results, we generated tTERT
with M194 substituted with leucine (M194L), thereby eliminat-
ing this cleavage site. The substitution was expected to cause an
increase of ?5 kDa in the size of the peptide that crosslinked to
the primer, which was the observed result (Fig. 1F; compare
lanes 2 and 4). A minor fragment was apparently produced by
incomplete digestion of the M194L tTERT at M235, generating
a radiolabeled fragment with the expected mass of the 14–411
peptide crosslinked to the primer.
To determine the location of the crosslinking site at single
amino acid resolution, we produced additional substitution
variants of tTERT and used them in extensive digestion assays.
First, we generated a tTERT in which the methionine residues
194 and 235 were substituted with leucine. This created a
methionine-free segment spanning amino acids 14–411 (Fig. 2A
Upper). Using the double mutant tTERT, we generated a series
of triple mutants in which nonconserved amino acids located at
various positions along the segment 145–191 were substituted
with methionine. These substitutions were designed to enable
cleavage of the 14–411 segment into two asymmetric fragments
resolvable by SDS/PAGE, only one of which should be radiola-
beled. The expected pairs of fragments that would result from
full digestion are shown in Fig. 2A Lower. SDS/PAGE of
crosslinking assays analyzed by extensive digestion revealed that
each tTERT variant, except W187M, produced one major
radiolabeled fragment and one or two additional minor radio-
labeled fragments. The major radiolabeled fragment produced
by the tTERT double mutant M194L/M235L (Fig. 2B, lane 1)
had a mass consistent with the expected 49.4-kDa segment
spanning amino acids 14–411 crosslinked to the DNA (Fig. 2A).
The major radiolabeled fragments produced by digestion of the
tTERT triple mutants with the substitutions E145M, L167M,
and K186M (Fig. 2B, lanes 2–4) migrated as expected if they
share a common C terminus at M411 and N-termini at the
positions specified by the third mutation (Fig. 2A, bold double-
arrowhead lines). The major radiolabeled fragments produced
by digestion of the tTERT triple mutants with the substitutions
Y188M, K189M, and N191M (Fig. 2B, lanes 6–8) migrated as
expected if they share a common N-terminal end at L14 and
C-termini at the positions specified by the third mutation (Fig.
2A, bold double-arrowhead lines). These data suggest that the
major crosslinking site maps between K186 and Y188 at a
tryptophan residue, W187. The crosslinking site W187 was
confirmed by assaying the tTERT triple mutant with the sub-
stitutions M194L/M235L and W187M, which did not generate a
major radiolabeled peptide (Fig. 2B, lane 5).
In addition to the major products described above (Fig. 2B,
marked by arrows), these assays generated distinct minor species
(Fig. 2B, marked by empty and filled arrowheads). The lengths
of the two species marked by empty arrowheads suggest that they
are partial digestion products. However, the single species
marked by a filled arrowhead could not be generated by partial
digestion. Therefore, it must represent a second crosslinking site
that maps in the segment spanning the amino acids 192–411. Of
thetTERTtriplemutantseriessharingM194L/M235L(Fig.2A),
this second crosslink was detected with mutants containing
W187M, Y188M, K189M, and N191M (Fig. 2B, lanes 5–8); in
the assays of mutants containing E145M, L167M, and K186M,
the second crosslink could not be detected because it resides
on the fragment that also includes the first crosslink.
We also studied crosslinks generated by using two additional
IdU-substituted primers 5?-GGG(IdU)TG-3? (IdU3) and 5?-
(IdU)TGGGGTT-3? (IdU8). To become radiolabeled subse-
quent to crosslinking, these two primers must align at the
beginningorendofthetemplate,respectively(Fig.2C).Asinthe
IdU5 primer used above, the IdU substitutions are positioned to
parallel the template-region residue A51. We found, using
extensive digestion assays (Fig. 2 D and E), that these primers
also crosslink to W187. Additionally, the primer IdU8 generated
aminorcrosslinkedspeciessimilartothatofIdU5(Fig.2E,filled
arrowhead), but we could not conclude whether IdU3 generated
this minor species as well. We also carried out crosslinking
assays, using primers with IdU substitutions at nucleotide resi-
dues aligned with the RNA template residues A50 and A45,
respectively. The first substitution gave a considerably lower
yield of crosslinking, and the second substitution gave an inde-
tectable crosslinking signal (data not shown).
Fig. 2.
mapping. (Upper) A complete map of the CNBr cleavage sites in the WT
enzyme is shown with enlargement of the region including the crosslinking
site. This scheme also shows the two methionines that have been substituted
with leucine in the double mutant M194L/M235L and the tryptophan residue
W187 that has been identified as the major crosslinking site. (Lower) Sche-
matized fragments produced by CNBr cleavage of additional tTERT variants,
whichincludedathirdsubstitutioninadditiontoM194L/M235L.Thenumbers
represent the fragments masses produced by complete cleavage with CNBr
and bold double arrows designate the major radiolabeled fragments that
were observed. The numbers in parentheses are the combined masses of
fragments and DNA, which are the actual species analyzed in the gel. (B)
SDS/PAGE of fragments from extensive digestion assays performed with the
tTERTmutantsshowninA,crosslinkedtotheprimerIdU5.Thearrowsindicate
themajorradiolabeledfragmentsproducedbycompletedigestionwithCNBr.
The empty arrowheads indicate the minor radiolabeled fragments generated
bypartialcleavage.Thefilledarrowheadpointsatthefragmentthatcontains
the second crosslinking site. (C) A scheme illustrating the template alignment
of the three substituted primers used for the assays shown in B, D, and E. The
stars indicate the active site. x designates the IdU substitution. (D) SDS/PAGE
of fragments from extensive digestion assays performed with the tTERT
mutantsshowninA,crosslinkedtotheprimerIdU3.(E)SDS/PAGEoffragments
from extensive digestion assays performed with the tTERT mutants shown in
A,crosslinkedtotheprimerIdU8.Thefilledarrowheadpointsatthefragment
containing the second crosslinking site.
Fine mapping of the crosslinks. (A) A scheme illustrating the fine
Romi et al.
PNAS ?
May 22, 2007 ?
vol. 104 ?
no. 21 ?
8793
BIOCHEMISTRY

Page 4

Telomerase Activity Assays of tTERT Molecules with Altered Amino
Acids at or near the Crosslinking Site. We next carried out telom-
erase activity assays of tTERT variants in W187 and adjacent
amino acids. We first synthesized tTERT with the substitution
W187A. This substituted tTERT was reconstituted into recom-
binant core enzyme and assayed for catalytic activity by direct
primer extension, using the unsubstituted version of the
crosslinked DNA 5?-GTTGGG-3?. In reactions containing
dTTP and radiolabeled dGTP, we found that the W187A
enzyme synthesized a profile of products that was indistinguish-
able from the comparably tagged WT enzyme (Fig. 3A, lanes 1
and 2). This remained true for reactions performed with dGTP
alone, with a longer primer (GTTGGG)3, with primers bearing
different 3? end permutations relative to the template and with
a broad range of primer concentrations [shown in supporting
information (SI) Fig. 5; additional data not shown]. Further-
more, no significant reduction in the radioactivity incorporated
into the longer extension products relative to the shorter prod-
ucts in the mutant versus the WT enzyme was observed (SI Fig.
5C). Thus, the W187A substitution does not appear to have a
substantial effect on the enzyme processivity. The various assays
comparing the activities of the WT enzyme and the W187A
variant were performed by using several different preparations
of each of the enzymes. The reason for the contrast between our
data and the severe catalytic defect reported for W187A in ref.
11 is not known.
To additionally characterize a potential primer binding defect
of the W187A enzyme, we determined the primer Km for
elongation. We used reactions containing radiolabeled dGTP
alone, so that product turnover was forced to occur after primer
extension by a single nucleotide. We found that primer Kmwas
little if at all affected by the W187A substitution (Fig. 3B). These
results could be due to the existence of additional interactions
between the primer and the enzyme that maintain the catalytic
activityoftheW187variant.ThefindingthattTERTW187isnot
essential for catalytic activity is consistent with our observation
that covalent crosslinking of W187 to primer still allowed primer
elongation.
Substitutions of tTERT were also made at residues flanking
W187. These substitutions were similarly assayed for catalytic
activity. SI Fig. 5 reveals that a substitution of the adjacent
tyrosine residue W188A did not substantially affect the activity
of the enzyme. We also tested tTERT alanine substitutions of
each of the four lysine residues, K183, K185, K186, and K189.
These individual substitutions, substitutions of pairs of these
lysine residues, or triple mutants of W187A, Y188A, and each
one of the lysine residues did not cause a substantial reduction
in the activity of the enzyme (data not shown). We did observe
a significant decrease in the activity of the substitution F158A
and of the triple substitution K185I/K186Q/K189A. Interest-
ingly, Jacobs et al. (11) observed a decrease in the activity of the
triple mutant K183A/K185A/K186A. Studies have reported that
substitution of Q168 strongly compromised catalytic activity (11,
17). We found that product synthesis by tTERT Q168A and
Q168E enzymes could be more readily detected if activity assays
contained high primer concentration (Fig. 3A, lanes 3–5). Assays
of the primer concentration-dependence of product synthesis for
theWTandtheQ168AandQ168Eenzymesrevealeda?10-fold
increase in primer Kmupon Q168A substitution and a ?4-fold
increase in primer Km upon Q168E substitution (Fig. 3C).
Importantly, among the various tTERT substitutions in this
region described in this work, only the Q168 substitutions were
found to have a specific impact on primer Km. These results
suggest that the Q168 side-chain has an influence on the binding
affinity of single-stranded DNA.
Discussion
We present here the first direct physical mapping at a single
amino acid resolution of DNA-TERT contacts in functional
telomerase complexes. Our crosslinking data clearly demon-
strated that in such complexes a close proximity, i.e., a chemical
interaction, occurs between tTERT W187 and a DNA primer
nucleotide that aligns with the 3? terminus of the RNA template
region. Furthermore, the same interaction was found to occur
with three different primers that had their 3? end aligned at
Fig. 3.
reconstituted in vitro with C-terminally FLAG-tagged WT or W187A tTERT, or
with untagged WT, Q168A or Q168E tTERT. (Upper) SDS/PAGE analysis of the
tTERT protein synthesis reactions used for activity assays is shown. Each tTERT
was expressed comparably. (Lower) Activity assay reactions performed in the
presence of 50 ?M GTTGGG primer, dGTP, and dTTP were analyzed by dena-
turing gel electrophoresis. (B) Double reciprocal plot of the rate of product
synthesis (v) versus primer concentration ([S]) for enzymes with C-terminally
FLAG-tagged WT and W187-substituted tTERTs. C-terminal FLAG-tagged
tTERTproteinswereusedfordirectcomparisonwithcrosslinkingassays,which
revealed covalent linkage of W187 to DNA. (C) Double reciprocal plot of the
rateofproductsynthesis(v)versusprimerconcentration([S])forenzymeswith
untagged WT and Q168-substituted tTERTs.
Activity assays of tTERT mutants. (A) Catalytic activity of telomerase
8794 ?
www.pnas.org?cgi?doi?10.1073?pnas.0703157104Romi et al.

Page 5

different positions along the template (Fig. 2 B–E). This appar-
ent uncoupling of the register of DNA–TERT interaction from
the register of DNA–template interaction is consistent with
results from our previous DNA footprinting and kinetic assays
(10, 18, 19).
In the absence of a high resolution crystal structure of a
complete telomerase enzyme, our data provide novel informa-
tion on the folding of the multidomain TERT in the enzyme.
Specifically, the data allow, for the first time, estimation of the
distances between TERT residues at the boundary of the TEN
domain and TERT residues in the RT domain active site (Fig. 4).
The precise spatial location of the DNA interaction with W187
relative to the RNA template is defined through the alignment
of the IdU substitution with the tTER residue A51. Thus, based
on the anticipated length and geometry of an RNA-DNA hybrid,
we can estimate the distance between W187 and the aspartic
acids in the active site of the telomerase. This distance varies
from ?17 to ?27 Å, depending on whether the primer 3? end
aligns with the template 3? end or 5? end, respectively (Fig. 4B).
The three primers used for the crosslinking assays (Fig. 2C)
simulate the sequential stages of product elongation by telom-
erase. The results of these assays suggest that during telomere
synthesis, the TEN and possibly the TRBD domains are dis-
placed relative to the active site. Such displacement could be
accommodated by structural flexibility of the region at the C
terminus of the TEN domain (11). The maximal accommodated
displacement would set an upper limit on the length of the
template-product hybrid, possibly accounting for the dissocia-
tion of base-pairs formed at the template 3? end before the
template 5? end can be copied (20–22). Curiously, mutation of
a tTER motif flanking the template 3? end led to the stalling of
product synthesis at the first template position proposed to
require unpairing of the template-product hybrid (23). This
region of tTER 3? of the template has been proposed to contact
the TEN domain directly (8), potentially on a surface distinct
from the groove for binding to single-stranded DNA (11).
Therefore, we speculate that TEN domain displacements could
coordinate the repositioning of both the template 3?-flanking
region of tTER and the template-adjacent DNA binding site of
tTERT relative to the active site during primer elongation.
Compared with previous assays of TERT–DNA interaction,
the strength of our procedure is that it ensures that the identified
contacts are functionally relevant. No crosslinked TERT–DNA
complexes would be labeled in our assays unless the primer is
properly aligned at the active site of an active enzyme. This
specificity is lacking in experiments using prelabeled DNA
molecules,whichareboundtovarioustelomerasecomplexesand
then crosslinked. The data obtained in such experiments could
represent binding of the DNA primers to inactive enzyme
molecules, or nonproductive binding to active molecules in a
manner irrelevant to the catalytic process. For similar reasons,
data on binding of DNA oligonucleotides to purified TERT do
not necessarily apply to functional enzyme complexes.
In addition to the contact with W187, the DNA must interact
with TERT by making other contacts, because the W187A
substitution does not substantially affect the activity of the
enzyme. One additional contact detected in our assays occurs
with a site located in a segment spanned by the amino acids
192–411. Additionally, we have shown that substitutions made in
the tTERT amino acid Q168 strongly affect the Kmfor primer
elongationbythereconstitutedcoreenzyme.Thesesubstitutions
may influence primer Kmdirectly through altering protein-DNA
contact, although we cannot preclude some impact of Q168
substitutions on local protein conformation. The recently solved
TEN domain structure places the Q168 side chain in a surface
groove predicted to bind single-stranded DNA (11). Interest-
ingly, W187 is located at the end of this groove (11).
Finally, we note that in the context of telomerase holoenzyme,
tTERT and/or other telomerase-associated proteins are likely to
provide additional sites for DNA interaction that contribute to
primer binding specificity and to processive elongation. Future
studies will be required to investigate the full spectrum of
telomerase interaction sites for single-stranded DNA.
Materials and Methods
Telomerase Reconstitution and Crosslinking. Telomerase core en-
zyme was reconstituted in rabbit reticulocyte lysate, as described
Fig. 4.
andtelomerases.MotifsTandT2arecommontoalltelomerases.MotifsCPandCP2arebestconservedamongTERTproteinsfromciliates.D618,D815,andD816
areconservedasparticacidresiduesattheactivesiteoftheenzyme,andW187isthesiteofcovalenttTERT-DNAcrosslinking.TEN,telomeraseessentialN-terminal
domain; TRBD, telomerase RNA binding domain; RT, reverse transcriptase domain. (B) A cartoon illustrating schematic models for folded structures of the
telomerase–DNA complexes that form at the beginning and at the end of a repeat synthesis cycle. In both, a primer nucleotide in the same template register
contacts tTERT W187. The blue asterisk indicates the position of the active site in the RT domain. The distance of ?17 Å was calculated based on the geometry
of the RNA-DNA duplex of 3 base pairs. The distance of ?27 Å was calculated for an RNA-DNA duplex of 8 bp. However, the actual distance could be slightly
larger because of unpairing of one or more nucleotides at the 3? end of the template.
Model of catalytically active telomerase complexes. (A) Linear scheme of tTERT with conserved motifs. Motifs 1, 2, A, B?, C, D, and E are common to RTs
Romi et al.
PNAS ?
May 22, 2007 ?
vol. 104 ?
no. 21 ?
8795
BIOCHEMISTRY