1. Biochemistry and Chemical Biology
Download icon

Control of the polyamine biosynthesis pathway by G2-quadruplexes

Research Article
Cite this article as: eLife 2018;7:e36362 doi: 10.7554/eLife.36362
5 figures and 2 additional files

Figures

Polyamines (PAs) and G-quadruplexes.

(a) PA biosynthesis pathway and summary of the roles/regulation of putative quadruplex sequences (PQS’s) on PA biosynthesis. Red: G2-rich motifs inhibit PA biosynthesis; Green: G2-rich motifs enhance PA biosynthesis; *: PA-sensing G2-rich motifs. (b) Guanine tetrad stabilized by a monovalent cation (M+) and intramolecular parallel G2- and G3-quadruplexes. Quadruplex loops - yellow. (c) G2-rich motifs in the UTRs of PSPs. PQS’s are denoted as the gene name with the UTR and PQS number (Qn; where n=≥1). Underlined guanines are those forming the most stable predicted G2-PQS from all possible PQS’s predicted by QGRS Mapper (version: Feb. 2014). Full length UTRs were utilized in the study in all cases except for the 5 UTRs of AMD1, AZIN1 and ODC1. *portion of the UTR used in this study for AMD1, AZIN1 and ODC1 (AMD1: 150-AUG; AZIN1: 720-241 and 95-AUG; ODC1: 301-AUG) (See Supplementary file 1 and Supplementary file 2 for details). ARG1, ARG2: arginase 1-2; ODC1: ornithine decarboxylase; SRM: spermidine synthase; SMS spermine synthase; AMD1: adenosylmethionine decarboxylase 1; AZIN1: antizyme Inhibitor 1; OAZ1-3: ornithine decarboxylase antizyme 1-3; SAT1: spermidine/spermine N1-acetyltransferase 1; SMOX: spermine oxidase; PAOX: polyamine oxidase.

https://doi.org/10.7554/eLife.36362.002
Figure 2 with 1 supplement
G2-PQS’s regulate luciferase reporter genes.

(a) Effect of PQS’s from PSP UTRs on Renilla luciferase activity in HeLa cells; PQS's ARG25Q1, AZIN15Q1, OAZ25Q1, OAZ25Q2, ODC15Q2, ODC15Q3, SMS5Q1, SAT15Q1, ARG23Q1, OAZ13Q1, OAZ33Q1 and SMS3Q1 show a statistically significant difference to their mutated controls (n = 3–7; *p≤0.05, **p≤0.01, ***p≤0.001); remaining PQS's are designated as inactive (n>/=2); error bars represent standard error (SE)). (b) Effect of functional PQS’s on Renilla mRNA levels in HeLa cells by qRT-PCR (n = 3–4, *p≤0.05; error bars represent standard error (SE)). (c) Properties of the functional G2-PQS’s. Underlined: predicted G-tetrad as in Figure 1c. Red, bold: long (>7 nt) loops. (d) Effects of G2-PQS’s on in vitro translation of Renilla luciferase in HeLa lysates (n = 3, *p≤0.05, **p≤0.01) (high variability of the SAT15Q1 was noted between different lysate batches). The effect of ARG23Q1 on in vitro transcription of Renilla luciferase. (See Figure 2—figure supplement 1). Error bars represent standard error (SE). Percentage changes discussed in the text are calculated as differences in normalized Renilla counts (x100).

https://doi.org/10.7554/eLife.36362.003
Figure 2—source data 1

Conservation of G2-PQS’s in PSP UTRs.

Assessment of PQS conservation using H-QGRS (Menendez et al., 2012), with the exception of SMS3Q1 and ARG23Q1 whose stability score is below that set by H-QGRS. SMS3Q1 and ARG23Q1 conservation was assessed through manual gene alignment. Conservation is dictated by factors such as composition, location and predicted stability. Human-mouse conservation was not observed for the reference NRAS, nor for ARG25Q1, ARG23Q1, OAZ25Q1, OAZ25Q2, SMS3Q1, OAZ13Q1 and OAZ33Q1, for which Human-primate conservation was assessed. For many PQS's, primate sequences are either not available or the UTRs are not defined (OAZ13Q1). Human-primate conservation was not observed for ARG25Q1, OAZ25Q1 and OAZ25Q2. Human-primate conservation has been reported for NRAS (Kumari et al., 2007).

https://doi.org/10.7554/eLife.36362.005
Figure 2—figure supplement 1
PAGE migration of long-looped G2-PQS’s in PSP UTRs upon in vitro transcription.

The effect of PQS destabilizing mutations (M) in AZIN15Q1, ARG23Q1, SMS3Q1, SAT15Q1 and SMS5Q1 on in vitro transcription of Renilla luciferase mRNA on replicate gels. Smeared RNA can be observed for SMS5Q1, possibly due to RNA aggregation ( left gel).

https://doi.org/10.7554/eLife.36362.004
Figure 3 with 4 supplements
Biophysical properties of G2-PQSs from PSPs: Gel migration, Thioflavin T fluorescence, UV (295 nm), CD.

All biophysical studies were performed in 100 mM K+. PAGE migration and Thioflavin T staining were performed at different migration times on distinct gels. Melting temps (TM 295) are shown in the plots; a K+-effect (1 mM and 100 mM K+) on TM 295 was found for ARG25Q1 (+18.8°C), OAZ25Q2 (+20.9°C) and SMS5Q1 (+18.4°C); mutated controls ARG25Q1M, OAZ25Q1M, SMS5Q1M, AZIN15QM1, AZIN15QM2, AZIN15QM3, did not show any (TM 295) melting transitions (data not shown and Figure 3—figure supplement 1; Figure 5—figure supplement 3). CD measurements were not performed for ARG23Q1 due to the presence of multiple RNA species. Error bars represent the standard error (SE) from two independent replicates. (Gels in this figure were cropped: full length gels are in Figure 3—figure supplement 3 and 4.

https://doi.org/10.7554/eLife.36362.006
Figure 3—figure supplement 1
UV melt-anneal profiles for mutant G2-PQS's.

UV melt-anneal profiles for NRASM, ARG25Q1M, ODC5Q2M, OAZ25Q1M, OAZ25Q2M, SMS5Q1M and ARG23Q1M at 295 nm (performed in the presence of 100 mM K+, UV performed in 1 mM K+ for NRAS). Two replicates of each biophysical experiment were performed. NRAS: UGUGGGAGGGGCGGGUCUGGG. NRASM: UGUAAAAGGGGCGGGUCUGGG.

https://doi.org/10.7554/eLife.36362.007
Figure 3—figure supplement 2
Additional biophysical data.

(a) Biophysical properties of the G2-PQS from SMS3Q1 and SAT5Q1 and mutants where relevant. Gel migration, Thioflavin T fluorescence, UV (295 nm), CD. All biophysical studies were performed in 100 mM K+. For the Thioflavin-T assay, the error bars represent the standard error (SE). (b) UV melt profile (295 nm) of AZIN15Q1 and CD trace of AZIN15Q1 and AZIN15Q1M. Performed in the presence of 100 mM K+. Two replicates of each biophysical experiment were performed.

https://doi.org/10.7554/eLife.36362.008
Figure 3—figure supplement 3
Full size migration gels from cropped images.

Full-length gels (shown as cropped images in Figure 3 and Figure 3—figure supplement 2).

https://doi.org/10.7554/eLife.36362.009
Figure 3—figure supplement 4
Full size thioflavin gels from cropped images.

Full-length gels (shown as cropped images in Figure 3 and Figure 3—figure supplement 2).

https://doi.org/10.7554/eLife.36362.010
Figure 4 with 7 supplements
G2-PQS in PSPs form quadruplexes in Hela cells and self-regulate.

(a) Effects of pyridostatin (PDS) on spermine and spermidine (n=3, *p≤0.05, **p≤0.01). (b) Effects of PDS on endogenous SMS protein at 6 and 24 h (n=3, *p≤0.05, **p≤0.01) (full length blots are in Figure 4—figure supplement 4); effects on SMS5Q1 (n=3, *p≤0.05) and SMS3Q1 (n=3-5, **p≤0.01, ***p≤0.001) wild-type and mutant reporter gene expression (UT: untreated); (c) Effects of PDS on endogenous AZIN1 protein at 24 h (n=3, *p≤0.05) (full length blots are in Figure 4—figure supplement 4); effects on AZIN15Q1 wild-type and mutant reporter gene expression (n=6, *p≤0.05). (d) Effect of PA supplementation on reporter activity from PQS’s in HeLa cells (n=3, *p≤0.05, **p≤0.01). (e) Effect of DMFO and APCHA on levels of spermine and spermidine in HeLa cells. Two independent replicates were performed. (f) Effect of PQS’s from PSPs on reporter activity in HeLa cells under PA depletion (0.5 mM DFMO, 100 µM APCHA, 6 days) followed by PA rescue (0.5 mM DFMO, 100 µM APCHA, 6 days, 100 µM PA, addition day 5). (n=3-5, *p≤0.05). Error bars represent standard error (SE) from at least two independent replicates. See Figure 4—figure supplements 5, 6 and 7.

https://doi.org/10.7554/eLife.36362.011
Figure 4—figure supplement 1
Cell viabilities after PDS treatments of HeLa cells.

Effects of the quadruplex stabilizing ligand Pyridostatin (PDS) on HeLa cell viability. The error represents the standard error (SE). (n = 3,**p0.01).

https://doi.org/10.7554/eLife.36362.012
Figure 4—figure supplement 2
Effects of PDS treatment upon spermine and spermidine levels in HeLa cells.

Example HPLC traces used for polyamine monitoring by HPLC at 229 and 198 nm, upon depletion of cellular polyamine levels following treatment with PDS for 24 hr (4, 16, 64 μM).

https://doi.org/10.7554/eLife.36362.013
Figure 4—figure supplement 3
Effects of PDS treatment on PQS reporter genes in HeLa cells: un-normalized data.

Effects of the quadruplex stabilizing ligand PDS on PQS-reporters in HeLa cells. Un-normalized Renilla data for the three reporters of Figure 4b and c after PDS treatment of cells.

https://doi.org/10.7554/eLife.36362.014
Figure 4—figure supplement 4
Effects of PDS treatment on endogenous polyamine synthesis proteins in HeLa cells.

Full-length blots.

https://doi.org/10.7554/eLife.36362.015
Figure 4—figure supplement 5
Polyamine addition and depletion reporter assays.

The error bars represent the standard error (SE) of at least two biological replicates. Normalized data.

https://doi.org/10.7554/eLife.36362.016
Figure 4—figure supplement 6
Polyamine addition and depletion reporter assays.

The error bars represent the standard error (SE) of at least two biological replicates. Un-normalized data.

https://doi.org/10.7554/eLife.36362.017
Figure 4—figure supplement 7
Polyamine depletion and spermine synthase inhibition.

Example HPLC traces used for polyamine monitoring by HPLC at 198, 229 and 254 nm, upon depletion of cellular polyamine levels following treatment with DFMO and APCHA. Confirmation of APCHA inhibition of spermine synthase under the conditions tested (putrescine or spermine addition).

https://doi.org/10.7554/eLife.36362.018
Figure 5 with 3 supplements
Structure of AZIN1wt.

(a) Derived model for the equilibrium between two possible hairpin conformers and the G2-quadruplex of AZIN1wt; K+ is expected to favor the G2-quadruplex; Na+ or Mg2+ are expected to favor the hairpin conformers, which are predicted by Mfold with free energies ΔG of −2.5 (left) and −1.8 kcal/mol (right). The mutants AZIN1neg and AZIN1pos were designed to stabilize the hairpin conformation and the G2-quadruplex, respectively: note the mutation of the first G2-tract at the 5’-end in AZIN1neg and the difference of the G2-quadruplex in the first loop from 5’-end (L1) for AZIN1pos. Sites of mutation are marked in red. (b) Overlay of 1H NMR spectra corresponding to the imino region of AZIN1wt (blue), AZIN1neg (black) and AZIN1pos (red) in 100 mM KCl (0.1 mM RNA). (c) Overlay of 1H NMR spectra corresponding to the imino region of AZINwt in 100 mM KCl (blue) and 200 mM KCl (red). (d) 1H NMR spectra after titration of spermine (Spm) to AZIN1wt in 100 mM KCl. Before addition of Spm (blue); after addition of Spm at RNA:Spm ratio of 1:1 (black) and 1:2 (red). Note the stronger decrease of the imino signals < 12 ppm.

https://doi.org/10.7554/eLife.36362.019
Figure 5—figure supplement 1
1H NMR spectra of AZIN1wt, NRAS, ODC15Q2, ARG25Q1, OAZ25Q1 and OAZ25Q2.

(a) Overlay of 1D 1H NMR spectra corresponding to the imino region of AZINwt in 100 mM KCl (blue) and 100 mM NaCl (red). (b) Overlay of 1D 1H NMR spectra corresponding to the imino region of AZINwt in 100 mM KCl (blue) and 2 mM MgCl2 (red). (c-d) 1D 1H NMR spectra corresponding to the imino region of NRAS and ODC15Q2 in 100 mM KCl. (e-g) 1D 1H NMR spectra and 1H-15N HSQC spectra corresponding to the imino region of ARG25Q1, OAZ25Q1 and OAZ25Q2 in 100 mM KCl. The imino protons observed at 11 ppm are bound to nitrogen atom with a chemical shift of about 145 ppm. The imino peaks at 12–13 ppm observed with OAZ25Q2 suggests that this RNA may form two alternative structures (G-quadruplex and a stem), similarly to what is observed with AZIN1wt. AZIN1pos: GGAGGCUUGGUGG; AZIN1neg: AAACCCAGACAUAGGCUUGGUGG; NRAS: GGGAGGGGCGGGUCUGGG; (0.03–0.5 mM RNA).

https://doi.org/10.7554/eLife.36362.020
Figure 5—figure supplement 2
1H 15N-HSQC and 1H 1H 2D NOESY of AZIN1wt.

(a) 1H-15N-HSQC spectrum corresponding to the imino region of AZIN1wt (in the presence of 10% 15N-labelled RNA) in 200 mM KCl. (b1H-15N HSQC spectrum corresponding to the imino region of AZIN1pos in 100 mM KCl. (c) 1H 1H 2D NOESY spectrum of AZIN1wt in 100 mM KCl. Red frames indicate the expanded regions. (0.1–0.5 mM RNA).

https://doi.org/10.7554/eLife.36362.021
Figure 5—figure supplement 3
Polyamine-related biophysical analysis of AZIN15Q1.

AZIN1 G2 quadruplex is induced by polyamines in vitro (a) Effect of spermine on UV melt profiles of (i) AZIN15Q1 (295 nm), (ii) control AZIN15Q1M3, (iii) control AZIN15Q1M2, (iv) control AZIN15Q1M (295 nm) and (v) AZIN15Q1 (260 nm) in the presence of 1 mM K+. Two replicates of each biophysical experiment were performed. (b) Effect of spermine for 1 hr on Thioflavin T fluorescence of AZIN15Q1, AZIN15Q1M, AZIN15Q1M2 and AZIN15Q1M3 in the presence of 1 mM K+ and 4.5 µM Thioflavin. The error bars represent the standard error (SE) from three independent replicates. AZIN15Q1M2: GGACCCAGACAUAGGCUUGGUAA; AZIN15Q1M3: AAACCCAGACAUAGGCUUGGUGG.

https://doi.org/10.7554/eLife.36362.022

Additional files

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)