We first compared the effects of various MPPs [FePP, ZnPP, SnPP and cobalt protoporphyrin (CoPP)] on cellular proliferation and DNA synthesis, (Fig. 1A-C) in Huh7 hepatoma and HEK293 embryonic kidney cells known to express telomerase. In contrast to other MPPs, ZnPP severely attenuated DNA synthesis (Fig. 1A) and depressed cellular proliferation greater than 50% at 48 hr treatment in both cell lines (Fig. 1B-C). Interestingly, ZnPP had only minor effects on proliferation in U20S cells, a line known to be telomerase negative (30) (Fig. 1D). As a positive control, BIBR 1532, a known mixed-type non-competitive inhibitor of telomerase (31) was tested in the same cell lines. BIBR showed greater propensity to decrease proliferation in telomerase-expressing rather than a telomerase negative line in accordance with earlier reports (Fig. 1D right chart), (32).
Consistent with effects on DNA synthesis and proliferation, ZnPP was the most effective MPP at inducing apoptosis (ca.: 50% at 10 uM) in TERT positive Huh7 or HEK293 cells in contrast to other MPPs evaluated (Fig. 1E). In support of the proliferation findings, ZnPP failed to have an increased effect on apoptosis in telomerase negative U2OS cells (Fig. 1E right panel).
Because of ZnPP actions on DNA synthesis, proliferation, and apoptosis in TERT positive cells, we determined ZnPP effects on TERT expression with western blots (WB) (Fig. 2A). As positive controls, we also evaluated other pro-proliferative proteins, β-catenin and cyclin D1 which have been closely linked to TERT expression and signaling (33, 34). ZnPP reduced expression of all three proteins by 8 -24hrs and by 48hr expression was nearly eliminated (Fig. 2A). Cells treated with various concentrations of ZnPP for 48hr (Fig. 2A, left middle panel) further confirmed these findings. In contrast, CoPP, FePP, or SnPP failed to significantly alter TERT, β-catenin, or cyclin D1 expression (Fig. 2A, right middle and bottom panels) consistent with their minimal effects on DNA synthesis and apoptosis (Fig. 1). The effects of ZnPP on TERT expression were apparent in different Huh 7 constructs and, interestingly, in a NS 5.15 HCV replicon, ZnPP promoted disappearance of both 120 KD telomerase monomer as well as the 45 KD C-terminal TERT fragment that we previously reported to be specific for HCV infected cells (Fig. 2B), (21).
We next evaluated the effects of MPPs on telomerase activity in cultured cells (Fig. 3A) as well as non-denatured cell lysates (Fig. 3B-E). In cultured cells incubated with ZnPP, telomerase activity was reduced in a dose-dependent fashion, (EC50 = 5.6–5.8 uM, upper and middle panels respectively) while SnPP or FePP, had none to mild effects (EC50 > 10 uM, either cell line) (Fig. 3A). The loss of telomerase activity with time of ZnPP treatment in the NS 5.15 HCV replicon (Fig. 3A, lower panel) reflected the disappearance of TERT seen in the WB (Fig. 2B).
The possibility that MPPs can directly inhibit telomerase activity in cellular extracts, similar to porphyrin quadruplex ligands, (6) was addressed next. Because of concerns that some quadruplex ligands inhibit Taq DNA polymerase in addition to telomerase, we assayed MPP inhibition at both steps of the TRAP procedure with a strategy similar to that of others (35). Using equivalent extracts but separate assays, MPP was either included in the telomerase RT extension step or the extension step was conducted without MPP and then MPP added only for the amplification steps with Taq DNA polymerase. To avoid further potential errors introduced by Realtime quantification, TRAP products were labeled with SYBR green, visualized on denaturing gels, and each lane quantified by absorbance measurements as described in the Methods. The latter step also ruled out the possibility that decreases in activity were artifactual due to fluorescence signal quenching by some MPPs (36). ZnPP was significantly more active (IC50 = 2.5 µM) than FePP and SnPP (both IC50 > 10.0 µM), (Figs. 3B-D respectively). All three MPPs had minimal effects on Tac polymerase during telomerase product extension and the slight inhibition of Taq polymerase seen for ZnPP was not directly concentration dependent. However, at least one MPP, CoPP, clearly inhibited Taq polymerase and could not be reliably assayed via TRAP assay (see Supplemental data, Fig. S1).
To confirm that ZnPP specifically inhibited telomerase as seen in the TRAP assays, direct telomere extension assays were conducted in the presence of α-32P-GTP. An IC50 of 2.7 µM for ZnPP obtained by direct extension assay was quite similar to the IC50 obtained with TRAP assay (2.5 µM) (Fig. 3E) Consequently, by two different assay procedures, ZnPP was observed to directly inhibit telomerase activity in cellular extracts and the IC50 values are roughly within a two fold range of the EC50 values for intact Huh7 and HEK293 cells (5 and 6 µM respectively). IC50 and EC50 values of the three MPPs obtained by different assay procedures are summarized in Table 2.
Table 2
Inhibition of telomerase by MPPs in direct telomere extension assays.
MPP | IC50 (µM) | R2 | Significant difference of IC50 [ZnPP < SnPP or FePP] |
a. Cellular lysates [Trapeze] |
ZnPP | 2.5 | .974 | |
SnPP | 12 | .962 | P < .001 |
FePP | > 20 | .911 | P < .001 |
b. α-32P-dGTP direct telomerase assay |
ZnPP | 2.7 | .904 | |
SnPP | > 16 | .645 | P < .001 |
FePP | > 20 | .777 | P < .001 |
c. Intact Huh7 cells [TRAP] |
| EC50 (uM) | | |
ZnPP | 5.4 | .941 | |
SnPP | 10.8 | .865 | P < .01 |
FePP | > 20 | .405 | P < .01 |
d. Intact HEK293 cells [TRAP] |
ZnPP | 6.4 | .928 | |
SnPP | > 20 | .076 | P < .001 |
FePP | > 20 | .085 | P < .001 |
e. TERT IP vs cellular lysate [TRAP] |
| IC50 (uM) | | IC50 of IP vs lysate |
ZnPP (IP) | 2.4 | .953 | NS |
ZnPP (lysate) | 1.8 | .963 | |
ZnPP, SnPP, and non-metal, “free” Lewis base protoporphyrins exhibit autofluorescence; (36) a property that has proven useful to study intracellular activities of MPPs such as nuclear localization and DNA or cellular adduct binding (12, 23, 37, 38). Other transition metal MPPs such as FePP or CoPP are inactive fluoroscopically because they have unfilled transition metal d orbitals that quench fluorescence emission. We investigated ZnPP binding to native, non-denatured telomerase-containing complexes after separation on large pore, (0.8%), agarose gels (27) (39). Initially, non-denaturing acrylamide gels were considered for these studies, however, we discovered that ZnPP labelled complexes would not enter the largest pore size possible, a result also noted by others (23, 37) (see Supplemental data Fig. 2).
Initially, increasing amounts of cellular extracts were incubated with varying amounts of ZnPP, then electrophoresed on large pore agarose gels. ZnPP was then visualized fluoroscopically using visible red wavelengths [608–632 nm Ex and 675–720 nm EM] or wide band UV (Fig. 4A upper and lower panels respectively). ZnPP bound to high molecular weight complexes in a concentration dependent manner and the complexes electrophoresed with a mobility just above thyroglobulin (670 kD), quite similar to sizes noted by us and others for TERT ribonuclear protein particles separated by glycerol gradient centrifugation (21, 40) and large pore agarose/acrylamide gels (39). Under these conditions, free ZnPP migrated slightly cathodal. While ZnPP binding was easily identified in cellular extracts, no binding was detectable in bulk protein incubations of BSA or IgG (Fig. 4B). ZnPP also labelled complexes in intact cells as determined by electrophoresis of extracts prepared after ZnPP incubation in culture (Fig. 4C).
To assess whether TERT is a component of the ZnPP-labelled cellular extracts we incubated lysates with anti-human TERT, β-catenin, or cyclin D1 antibodies or non-specific IgG antibodies prior to electrophoresis and looked for upward mobility shift after electrophoresis (Fig. 4D). Only anti-TERT antibody led to a significant upward mobility shift of ZnPP labeled complexes from either cell type, suggesting that TERT is indeed a component of the large complexes. No mobility shift was noted for the other antibodies tested suggesting ZnPP specifically labelled TERT complexes. Note that both cyclin D1 and β-catenin would be expected to be components of large molecular complexes in non-denatured cellular lysates (41, 42). Further characterization of the ZnPP binding complexes as to protein and DNA composition, and investigation of SnPP binding is presented in the online supplemental data (Fig. S2).
ZnPP labelled complexes from Huh7 cells were next blotted onto nitrocellulose by capillary diffusion (conditions determined empirically, see Supplementary data, Fig. S3) and probed with specific anti-TERT or anti-dyskerin antibodies, the latter a positive control for telomerase holoenzyme (Fig. 4E). Both TERT and dyskerin were easily identified in the high molecular weight complexes binding ZnPP (Fig. 4E). Interestingly, cellular lysates showed more immunoreactive TERT when incubated with ZnPP prior to electrophoresis suggesting a protective effect of ZnPP on TERT in the extracts.
Immunoprecipitation using TERT -specific antibodies further confirmed that TERT is a component of ZnPP labelled complexes. Immunoprecipitates (IP) were evaluated on native agarose as well as denaturing SDS gels and WB (Fig. 5A, left and right panels respectively). ZnPP bound the anti-TERT IP complexes intensely and IP had increased mobility as compared to non-specific IP complexes or no antibody control (Fig. 5A, left panel). As expected, the anti-TERT IP analyzed on WB (Fig. 5A right panel) showed increased TERT as compared to IP from non-specific or no antibody controls. A weaker TERT band (relative to exposure time) was also identified in the crude cellular lysate. When assayed by TRAP assay, IP TERT complexes and unpurified enzyme showed similar IC50 values with ZnPP, (3.1 vs 2.2 uM respectively), (Fig. 5B). These measurements were also close to the IC50 observed for overexpressed enzyme, (2.7 uM), when assayed by direct α-P32-GTP extension assay (Fig. 3E), (Table 2).
To investigate whether nucleic acids are components of the ZnPP labelled complexes; IP or crude lysates, were digested with DNase I or RNase A prior to labelling with ZnPP (Fig. 5C, upper panel). Digestion of extracts with DNase I elicited minimal changes in the mobility of ZnPP labelled complexes, however, a significant upward shift was seen when extracts were digested with RNase A. Furthermore, there was a marked increase in unbound ZnPP after nuclease, most dramatic with RNase A – treated IPs, (arrow, Fig. 5C, upper panel) suggesting that ZnPP most likely binds to a ribonuclear protein complex and the binding site is at least partially disrupted with RNase digestion. Next, nuclease digested, ZnPP labelled complexes were blotted onto nitrocellulose and reacted with anti-TERT antibodies. These experiments showed that RNase A digestion severely diminished the amount of immunoreactive TERT in the labelled complexes, most notably for IP complexes (Fig. 5C, lower panel). These findings strengthened the conclusion that ZnPP binds to high molecular weight telomerase complexes. In addition to providing a telomere template, TERC is an important structural component of telomerase holoenzyme and RNase A digestion is known to completely disrupt ribonuclear complex structure and release TERT (43).
ZnPP binding to cellular structures in situ was probed by confocal immunofluorescence microscopy. While telomere sequences exist throughout the mammalian chromosome, it is known that telomerase holoenzyme only associates with telomeres at DNA replication during S phase (44). Consequently, we compared ZnPP localization in synchronized S-phase cells as compared to metaphase chromosomes which contain prominent telomere ends without telomerase. The percentages of cells in S phase were determined temporally with flow cytometry after double thymidine block and extracts were monitored on immunoblots with Cyclin A2 staining to determine optimal times for study, (Fig. S4 Supplementary data). Telomere sites were labelled with telomere sequence specific fluorescent probe (PNA TEIC-Alexa488 F1004) and TERT was localized with specific antibodies (Fig. 6).
First, we looked at ZnPP co-localization with the telomere probe. While ZnPP clearly localized with telomeres in S phase cells, (Fig. 6A left panels), it did not label the prominent telomeres on metaphase chromosome tips, (Fig. 6A, right panels) which are devoid of holoenzyme. Next, we investigated whether ZnPP would co-localize with TERT in S phase as compared to unsynchronized Huh7 cells. In S phase cells, TERT co-localized with ZnPP in the nucleus and at some cytoplasmic sites (Fig. 6B left panels). As we reported previously, TERT is sparsely present in the nucleus of unsynchronized, log phase Huh7 cells, but is chiefly found at perinuclear sites which co-localize with mitochondria (21). Interestingly, even perinuclear TERT, likely lacking telomeric DNA, showed avid TERT- ZnPP co-localization (Fig. 6B, right panels). Collectively, these data indicate that ZnPP can bind to telomerase complexes and/or associated components. While telomeric DNA does not appear to be a primary binding site of ZnPP per se, at least at prominent telomeres on metaphase chromosomes, the specific sites of interaction in the telomerase holoenzyme remain to be determined.