(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



Insights into POT1 structural dynamics revealed by cryo-EM

['Emmanuel W. Smith', 'Ntu Institute Of Structural Biology', 'Nanyang Technological University', 'Singapore', 'Simon Lattmann', 'Zhehui Barry Liu', 'Bilal Ahsan', 'Daniela Rhodes', 'School Of Biological Sciences']

Date: 2022-02 

Telomeres are protein-DNA complexes that protect the ends of linear eukaryotic chromosomes. Mammalian telomeric DNA consists of 5′-(TTAGGG)n-3′ double-stranded repeats, followed by up to several hundred bases of a 3′ single-stranded G-rich overhang. The G-rich overhang is bound by the shelterin component POT1 which interacts with TPP1, the component involved in telomerase recruitment. A previously published crystal structure of the POT1 N-terminal half bound to the high affinity telomeric ligand 5′-TTAGGGTTAG-3′ showed that the first six nucleotides, TTAGGG, are bound by the OB1 fold, while the adjacent OB2 binds the last four, TTAG. Here, we report two cryo-EM structures of full-length POT1 bound by the POT1-binding domain of TPP1. The structures differ in the relative orientation of the POT1 OB1 and OB2, suggesting that these two DNA-binding OB folds take up alternative conformations. Supporting DNA binding studies using telomeric ligands in which the OB1 and OB2 binding sites were spaced apart, show that POT1 binds with similar affinities to spaced or contiguous binding sites, suggesting plasticity in DNA binding and a role for the alternative conformations observed. A likely explanation is that the structural flexibility of POT1 enhances binding to the tandemly arranged telomeric repeats and hence increases telomere protection.

Currently, information on the three-dimensional structure of the shelterin complex, or any of the isolated full-length protein subunits is sorely lacking. Structural information is limited to crystal structures of some of the sub-domains of the shelterin subunits, or complexes of protein-protein interfaces ( Fig 1C ) [ 23 – 32 ]. Here, we report on our efforts to determine the structure of the shelterin sub-complex POT1-TPP1-TIN2(1–354) by cryo-EM. Our attempts to obtain the structure of the full shelterin complex (POT1-TPP1-TIN2(1–354)-TRF2-RAP1) were unsuccessful, likely due to high conformational heterogeneity. We have reconstructed two 3D maps: One at 7.9 Å resolution into which we can fit the full-length POT1 protein complexed to the PBD of TPP1. While we could unambiguously fit the structures of the POT1 OB1 and the POT1 OB3/HJR domain bound to the PBD of TPP1 into the cryo-EM map, densities for the remaining TPP1 and TIN2 proteins are missing. The second map at 9.6 Å resolution displays an alternative conformation for POT1, in which OB1 and OB2 occupy densities that are spaced further apart from each other, suggesting POT1 has a dynamic structure allowing the two OB folds to take up an alternative conformation. Finally, to test the hypothesis that an extended conformation of the two OB folds could affect DNA recognition and that binding was not exclusively restricted to the minimal continuous binding sequence TTAGGGTTAG, we performed DNA binding studies using telomeric ssDNA ligands in which spacers were introduced between the half-sites for OB1 and OB2. We find that ligands with spaced half-sites indeed bind with no significant loss of affinity. We speculate that the alternative orientations we observe for the POT1 OB folds might have important implications for the recognition of telomeric tandem repeats and telomere protection.

The telomeric decamer sequence 5′-TTAGGGTTAG-3′ is the minimal tight binding sequence for POT1 with a reported apparent K d of 10 nM [ 25 ]. A crystal structure (PDB ID: 1XJV) of the N-terminal half of POT1 bound by this decamer sequence revealed that the two OB folds (OB1 and OB2) pack together forming a cavity along their axis where the single-stranded DNA ligand binds ( Fig 1C ) [ 25 ]. The interactions with the DNA strand are orchestrated mainly through the stacking of bases with hydrophobic residues, while the phosphate backbone is solvent exposed [ 25 ]. The structure also revealed that OB1 binds to a half-site consisting of the first six nucleotides (TTAGGG), while the adjacent OB2 binds to a half-site consisting of the last four nucleotides (TTAG) of the binding-site [ 25 ]. It was also reported that dT2, dA3, dG4, and dG5 of the ssDNA ligand contribute the most to binding, demonstrating that most of the DNA-binding affinity stems from the OB1 fold [ 25 ].

( A ) Cartoon representation of the full shelterin complex POT1-TPP1-TIN2-TRF2(2×)-RAP1(2×) bound to a telomeric double-stand/singe-stand junction. ( B ) Interaction map for the shelterin subcomplex POT1-TPP1-TIN2. The OB3/HJR domain of POT1 (orange) interacts with the PDB of TPP1 (blue), while the TBM of TPP1 (blue) interacts with TIN2 through the TRFH domain of TIN2 (green). TIN2 also has a TBM and a DC motif. ( C ) Published crystal structures of domains of the POT1-TPP1-TIN2 shelterin subcomplex: POT1 (OB1 & OB2) (yellow) bound to the minimal telomeric sequence TTAGGGTTAG (red) (PDB ID: 1XJV) (top left), POT1 (OB3 & HJRD) (yellow) bound by the PBD of TPP1 (E266-L326) (blue) (PDB ID: 5UN7, 5H65) (top right), TPP1 OB domain (blue) (PDB ID: 2I46) (bottom left), and TIN2(green)-TPP1(blue)-TRF2(grey) interface complex (PDB ID: 5XYF) (bottom right).

Shelterin is a multiprotein complex that binds mammalian telomeres forming a protective capping structure [ 11 – 13 ]. It consists of the protection of telomeres protein 1 (POT1), the POT1 interacting factor TPP1, the telomeric repeat-binding factors 1 and 2 (TRF1 and TRF2) as well as RAP1 and TIN2 ( Fig 1A ) [ 11 , 13 , 14 ]. POT1 on the one end of the complex binds to the telomeric single-strand G-overhang, while homodimers TRF1 or TRF2 at the other end of the complex bind to the telomeric double-stranded repeats, and together, these proteins anchor the full shelterin complex onto the single-strand/double-strand junction ( Fig 1A ) [ 11 , 15 ]. Additionally, POT1 complexed to TPP1 binds independently and sequence-specifically to multiple sites, coating the telomeric G-overhang [ 16 , 17 ]. Finally, shelterin or sub-complexes containing at a minimum only the POT1 and TPP1 subunits can recruit telomerase through the TEL patch domain on TPP1, and hence regulate telomere elongation and prevent severe telomere instability [ 18 – 21 ].

Telomeres are protein-DNA complexes that protect chromosomes from degradation by preventing end-to-end fusion and countering the end replication problem [ 1 – 3 ]. Mammalian telomeric DNA consists of 5′-(TTAGGG)n-3′ double-stranded repeats, followed by up to several hundred bases of single-stranded telomeric repeats forming a 3′ G-rich overhang (G-overhang) [ 1 – 4 ]. Telomeres shorten with every cell division, and while this progressive shortening is generally considered a hallmark of aging, in stem and cancer cells it can be countered by a dynamic interplay between the telomerase enzyme that synthesizes telomeric repeats and the telomere-binding protein complexes shelterin and CST (CTC1, STN1, TEN1) [ 5 – 8 ]. Consequently, telomeres and the protein complexes that regulate their structure or stability are intensively studied in attempts to develop therapies for cancer and age-related disease [ 4 , 9 , 10 ].

The experimental binding conditions of the WEMSA were quasi-identical to the EMSA except that the total concentration of the telomeric DNA oligonucleotide was raised from 10 pM to 10 or 20 nM by supplementing the reactions with unlabeled telomeric DNA of identical sequence. Subsequent to native PAGE fractionation, the nucleic acid–protein complexes were electroblotted (20 V, 120 min, 4° C) onto a PVDF membrane under non-denaturing (0.5× TB) conditions. The PVDF membrane was soaked in 10% acetic acid for 15 min and subsequently air-dried to denature and fix the proteins to the membrane. The nucleic acid–protein complexes were firstly detected by phosphorimaging. In a second step, the proteins bound to the membrane were identified by immunoblotting. Antibody binding was visualised using WesternBright enhanced chemiluminescence HRP substrate (Advanstra) and detected by a CCD camera imager (ChemiDoc Touch, Bio-Rad) ( S7A Fig ). Tight specificity of the TIN2 (ab197894, Abcam), TPP1 (ab112050, Abcam) and POT1 (ab124784, Abcam) antibodies for their respective antigen and the absence of cross-reactivity with the other components of the complex was verified beforehand ( S7B Fig ).

Where y is the fraction of unbound DNA at equilibrium, F min is the minimal unbound DNA fraction, F max is the maximal unbound DNA fraction, K d the equilibrium dissociation constant, x is the protein complex concentration at equilibrium and H is the Hill coefficient. Each experiment was repeated at least three independent times ( S6 Fig ).

Purified recombinant POT1-TPP1-TIN2(1–354) protein complex at concentrations from 100 nM to 10 pM (10 1/3 -fold serial dilution) was incubated with 10 pM 5′- 32 P-labelled telomeric DNA oligonucleotides in a total volume of 22 μl EMSA buffer (50 mM Tris-HCl (pH 8.0), 150 mM KCl, 4 mM MgCl 2 , 0.1% (v/v) Tween-20, 100 μg/ml BSA, 10% (v/v) glycerol, 0.5 ng/μl poly(dI:dC), 1 mM DTT and 1 μM random sequence carrier ssDNA). The reactions were equilibrated at 30° C for 30 min prior to electrophoresis. From each reaction, a 10-μl aliquot of the nucleic acid–protein complex was fractionated (200 V, 35 min) on a pre-electrophoresed non-denaturing (0.5× TB buffer) 4% polyacrylamide gel (37.5:1 mono:bis ratio). Gels were subsequently dried onto a positively charged Hybond N+ nylon membrane (GE Healthcare) and phosphorimaged (Typhoon FLA 7000). The signal intensity of the free DNA was quantified with ImageQuant TL software (Nonlinear Dynamics). The binding affinity (K d ) of the POT1-TPP1-TIN2(1–354) complex for each DNA oligonucleotide was determined by fitting 14 experimental data points to a four-parameter logistic equation using the Levenberg-Marquardt algorithm:

( A ) Representative micrograph. A total of 6,818 micrographs were collected and processed. ( B ) Representative 2D class averages from RELION reference-free 2D classifications. ( C ) Data processing workflow for cryo-EM map reconstruction. ( D ) FSC curves for the two conformations observed, resulting from gold standard 3D refinement. Resolution was determined with a 0.143 FSC threshold. ( E ) Local resolution estimated by Monores and angular distribution analyzed from RELION. Most parts of the two conformational EM structures are resolved in the resolution range of 6–10 Å.

Raw movies for the data from the 0° and 30° stage tilt were combined and pre-processed using Warp’s [ 36 ] automated pre-processing pipeline. The combined movies were motion-corrected with a spatial resolution of 5 × 5 for local motion compensation. Contrast transfer function (CTF) parameters were estimated with 9 × 9 spatial resolution for correcting local differences in defocus. Bad micrographs were excluded at this stage based on manual inspection and quality filters, and a curated set of 5,037 micrographs ( Fig 2A ) was retained. Particles were auto-picked with a retrained BoxNet model based on ~1,000 manually picked particles. A total of 2,103,136 particles were extracted with a box size of 200 × 200 pixels, imported to cryoSPARC v2 [ 37 ], and subjected to two rounds of 2D classification ( Fig 2B ). The selected subset of 620,136 particles was used to generate an initial 3D reference model in cryoSPARC v2, and used for downstream image-processing in RELION. Following the first round of 3D classification with 5 classes inside RELION-3.0 [ 38 ], two conformations with a dominant population and distinctive features were observed and selected ( Fig 2C ) for another round of 3D classification with local angular searches to reduce conformational heterogeneity. A subset 3D class of conformation 1 that consists of 41,547 particles was selected and subjected to a 3D auto-refinement job, which refined to a 7.9 Å resolution EM map ( Fig 2C ). Likewise, a similar 3D classification and refinement step was performed on conformation 2 and a 9.6 Å resolution reconstruction was obtained from a subset of 32,488 particles ( Fig 2C ). A soft mask was created for both of the two conformational maps and used for post-processing in RELION. The resolutions reported were calculated based on the gold-standard Fourier shell correlation (FSC) at 0.143 criteria ( Fig 2D ) [ 39 ]. Local resolution was calculated with MonoRes [ 40 ] in Scipion [ 41 ] with refined half maps and masks created from RELION as input ( Fig 2E ).

Frozen 0.7 mg/ml aliquots of POT1-TPP1-TIN2(1–354) were thawed and dialyzed overnight at 4°C in a dialysis cap against a glycerol-free buffer (200 mM NaCl, 20 mM Hepes pH 7.8, 1 mM 2-mercaptoethanol). The sample was further diluted to 0.05 mg/ml, and 3 μl drops were applied onto 1.2/1.3 UltrAuFoil Grids inside a Vitrobot Mark IV (Thermo Fisher Scientific) chamber maintained at 100% humidity and 4°C. The grids were blotted for 2 seconds at 0 blot force before plunge freezing into liquid ethane and finally transferred to a Titan Krios Transmission Electron Microscope (Thermo Fisher Scientific) operated at 300 keV. Electron micrographs with a calibrated object pixel size of 1.1 Å /pixel were recorded using the EPU software (v 2, Thermo Fisher Scientific, USA) on a K2 direct electron detector (Gatan) operated in counting mode. During data collection, a 20 eV slit was inserted in the Bioquantum 969 energy filter (Gatan), and the zero-loss peak was aligned every two hours. The data acquisition illumination provided an electron flux of 6 electrons/pixel/sec, and the exposure time was accordingly set to provide fluence of 50 electrons/Å 2 while images were collected in movie-mode and fractionated into 40 frames. All images were recorded at -0.5 μm defocus with a Volta Phase Plate inserted at the back focal plane of the objective lens to improve image contrast. The position of the Volta Phase Plate was set to change every 70 images during data acquisition. On-plane conditions, condenser and objective lens astigmatism, and coma-free alignments were carried out before data acquisition on the carbon area of a Quantifoil grid. Due to severe orientation bias exhibited by the particles, data were collected both at 0° and 30° stage tilt.

Analysis of the elution profile for the telomeric ssDNA ligand (sstelo64) alone or combined suggests the complex is binding ( S3 Fig ) and was also verified by (W)EMSA ( S6 and S7 Figs). Additionally, TIN2(1–354) appears to effectively bridge the POT1-TPP1 and TRF2-RAP1 components forming a larger complex as indicated by the position of the elution profile when combining the POT1-TPP1-TIN2(1–354) and the TRF2-RAP1 complexes ( S2B Fig ). Lastly, titrating the telomerase ligand (telo666, S1B Fig ) with increasing concentrations of the POT1-TPP1-TIN2(1–354) shows a marked dose-dependent augmentation in telomerase processivity ( S2C Fig ), consistent with previous observations [ 27 ]. Similarly, a titration with the full shelterin complex also manifested an increase in telomerase processivity, although with a greater stimulatory effect at low concentrations in comparison to the POT1-TPP1-TIN2(1–354) complex alone. As a control, the TRF2-RAP1 complex alone had no effect on telomerase processivity. In summary, the purified full shelterin complex as well as the POT1-TPP1-TIN2(1–354) shelterin sub-complex we have produced appear to be fully functional in shelterin assembly and DNA binding, and are proficient in telomerase recruitment to the G-overhang for telomere repeat synthesis. Therefore, both complexes were judged suitable for cryo-EM structural studies.

To establish that the purified POT1-TPP1-TIN2(1–354) sub-complex produced by over expression was fully functional, we analyzed its ability to assemble into a fuller shelterin complex ( S2B Fig ), its DNA binding properties ( S3 , S6 and S7 Figs), as well as its effect on the enzymatic activity of the telomerase enzyme ( S2C Fig ).

The human telomerase was overexpressed and purified as previously described [ 35 ]. Purified recombinant shelterin protein complexes (POT1-TPP1-TIN2(1–354), TRF2-RAP1 or POT1-TPP1-TIN2(1–354)-TRF2-RAP1) at concentrations from 100 to 5 nM (20 1/7 -fold serial dilution) were incubated with 10 nM telo666 telomeric ligand ( S1B Fig ) in a total volume of 10 μl telomerase assay buffer [50 mM Tris-HCl (pH 8.0), 50 mM KCl, 1 mM MgCl 2 , 4 mM 2-mercaptoethanol, 500 μM each of dATP and dTTP, 20 μM dGTP and 20 nM [α- 32 P]dGTP (6000 Ci/mmol, PerkinElmer)]. The binding reactions were pre-equilibrated at 30°C for 30 min. Telomeric DNA extension reactions were initiated by the addition of purified recombinant telomerase and incubated at 30° C for an extra 30 min. Reactions were terminated by addition of an equal volume of deionised formamide. After addition of proteinase K (0.5 mg/ml), the reaction mixtures were further incubated at 37° C for 20 min. The reaction products were heat-denatured for 10 min at 95°C and one-half of the reaction volume was resolved (17 W, 65 min) on a pre-electrophoresed (17 W, 60 min) denaturing (7 M urea, 0.5× TBE) 8% polyacrylamide (19:1 mono:bis ratio) gel. Gels were subsequently dried onto a positively charged Hybond N+ nylon membrane (GE Healthcare) and phosphorimaged (Typhoon FLA 7000).

Analytical size exclusion chromatography was used to assess the binding of telomeric ssDNA ligands to the POT1-TPP1-TIN2(1–354) complex. It was performed using a Superose 6 Increase 10/300 GL column pre-equilibrated with a 300 mM NaCl, 20 mM Hepes pH 7.5, and 1 mM TCEP buffer. 600 pmol of purified recombinant POT1-TPP1-TIN2(1–354) was injected onto the SEC column alone, or after incubation (30 min, 4° C) with 1200 pmol of the telomeric ssDNA ligand GGTTAGGGTTAG (sstelo64). For comparison, 600 pmol and 1200 pmol of the telomeric ssDNA ligand were injected separately, and the elution profiles were superimposed and analyzed ( S3 Fig ). To test the capacity of the TIN2(1–354) variant to bridge TRF2/RAP1 and reconstitute the full shelterin complex (POT1-TPP1-TIN2(1–354)-TRF2(2×)-RAP1(2×)), 200 pmol POT1-TPP1-TIN2(1–354) or 400 pmol of TRF2-RAP1 were first injected alone onto a Superdex 200 Increase 10/300 GL column pre-equilibrated with a 300 mM NaCl, 20 mM Hepes pH 7.5, and 1 mM TCEP buffer. Subsequently, 200 pmol POT1-TPP1-TIN2(1–354) was combined with 400 pmol of TRF2-RAP1 and equilibrated (30 min, 4° C) prior to being injected onto the SEC column. The chromatograms from the three separate runs were then superimposed and analyzed for complex formation ( S2B Fig ).

The fully assembled shelterin complex (POT1, TPP1, TIN2, TRF2 and RAP1) was produced by independently overexpressing POT1-TPP1-TIN2 and TRF2-RAP1 as two separate sub-complexes and purifying each sub-complex to high purity ( S2A Fig ). For the POT1-TPP1-TIN2 shelterin sub-complex, a construct containing full-length POT1, full-length TPP1 and a C-terminally truncated allele of TIN2(1–354) was chosen as it yielded the highest expression of the complex, and because the C-terminus of TIN2 contains no known protein-interacting domains [ 34 ]. The full shelterin complex was then assembled by combining the two purified sub-complexes, followed by fractionation by size exclusion chromatography.

For TRF2-RAP1, an Sf9 cell pellet was resuspended in lysis buffer (300 mM NaCl, 20 mM Hepes pH 7.5, 2 mM MgCl 2 , 0.5 mM TCEP, DNase, 1 mM Benzamidine, 0.1% (v/v) Triton X-100 and 1 mM PMSF) and sonicated (40 amplitude, 3 s on/5 s off, 10 min), before being centrifuged at 25,000g for 1 hour. The clarified cell lysate was filtered and loaded onto a 1-ml StrepTrap HP column at 1 ml/min flow rate. The immobilized protein was then washed (300 mM NaCl, 20 mM Hepes pH 7.5, and 0.5 mM TCEP) followed by a high-salt wash (1 M NaCl, 20 mM Hepes pH 7.5, and 0.5 mM TCEP) and eluted with elution buffer (300 mM NaCl, 20 mM Hepes pH 7.5, 0.5 mM TCEP, and 5 mM D-Desthiobiotin). The eluted sample was incubated overnight in the presence of 0.5 mg TEV protease and was then concentrated to 100 μl and injected onto a Superdex 200 Increase 10/300 GL column pre-equilibrated with the Size Exclusion Buffer (300 mM NaCl, 20 mM Hepes pH 7.5, 10% (v/v) Glycerol, 0.5 mM TCEP). The TRF2-RAP1 complex eluted around 11.3 ml and the purity as assessed by SDS-PAGE was found to be >95% ( S2A Fig ). The fractions corresponding to the peak were pooled and the sample was dispensed into 50-μl aliquots (1.25 mg/ml), flash frozen in liquid nitrogen and stored at -80°C for future use.

A pellet of Sf9 cells expressing the POT1-TPP1-TIN2(1–354) construct was resuspended in lysis buffer (300 mM NaCl, 20 mM Hepes pH 7.5, 25 mM Imidazole, 5 mM MgCl 2 , 1 mM TCEP, DNase, 1× cOmplete™- EDTA-free Protease Inhibitor Cocktail, 0.1% (v/v) Triton X-100, 1 mM PMSF) and sonicated (40 amplitude, 3 s on/5 s off, 10 min), before being centrifuged at 25,000g for 1 hour. The clarified cell lysate was filtered and loaded onto a 1-ml HisTrap HP column at 1 ml/min flow rate. A salt wash was then performed with buffer containing 1 M NaCl, 20 mM Hepes pH 7.5, 25 mM imidazole, and 1 mM TCEP, followed by a wash with buffer containing 300 mM NaCl, 20 mM Hepes pH 7.5, 74 mM imidazole, and 1 mM TCEP. Elution was performed over a 10-column volume gradient to 100% Elution Buffer (300 mM NaCl, 20 mM Hepes pH 7.5, 750 mM imidazole, 1 mM TCEP) at 1 ml/min flow rate. The eluted sample was then supplemented with 0.5 mg TEV protease and dialyzed overnight in dialysis buffer (300 mM NaCl, 20 mM Hepes pH 7.5, 25 mM imidazole, 5 mM MgCl 2 , 1 mM TCEP) to reduce the imidazole concentration. The dialyzed sample was reloaded onto a HisTrap HP 1-ml column to remove the uncleaved protein, the TEV protease, and contaminants. The flowthrough was then concentrated to 100 μl and injected onto a Superdex 200 Increase 10/300 GL column pre-equilibrated with the Size Exclusion Buffer (300 mM NaCl, 20 mM Hepes pH 7.5, 10% (v/v) glycerol, 1 mM TCEP). The complex eluted around 13.5 ml and the purity as assessed by SDS-PAGE was found to be >95% ( S2A Fig ). The fractions corresponding to the peak were pooled and the complex dispensed into 100-μl aliquots (0.7 mg/ml), flash frozen in liquid nitrogen and stored at -80°C for future use.

For expression of the TRF2 and RAP1 shelterin components, untagged full-length RAP1 (1–399) and full-length TRF2 (1–542) fused to an N-terminal Twin-Strep-tag and a TEV protease cleavage site were cloned into the pFL plasmid through Gibson assembly. The bicistronic construct was then inserted into the MultiBac bacmid and the remaining steps were implemented similarly as for the POT1-TPP1-TIN2 constructs.

Full-length human POT1 (1–634) fused to a TEV protease cleavage site and an N-terminal His-10 tag was cloned into the pFL acceptor plasmid through Gibson assembly [ 33 ] along with an N-terminal Strep-tag II tagged variant of TPP1 (full-length (1–544), N-terminal truncation variant (87–544), or both N-terminal and C-terminal truncation variant (87–334)) also fused to a TEV protease cleavage site. Similarly, either full-length (1–451) or a C-terminal truncation variant (1–354) of TIN2 was cloned into the donor pUCDM plasmid. Cre-LoxP fusion reaction was used to recombine the acceptor and donor plasmids and the correct constructs were selected in the presence of ampicillin and chloramphenicol. The chosen multicistronic constructs were subsequently subcloned into the MultiBac bacmid and further selected by blue-white screening. Sf9 insect cells were transfected with the bacmid DNA and used to amplify the virus. The viral titer was determined by the end-point dilution assay. For expression of POT1-TPP1-TIN2 constructs, 700 ml Sf9 cell cultures (at 3×10 6 cells/ml) were infected (MOI = 1.5) and cultured for 3 days in suspension at 27° C. Cells were harvested by centrifugation and pellets stored at -80°C for future use.

All DNA oligonucleotides ( S1B Fig ) were purchased from Integrated DNA Technology. For EMSA and WEMSA experiments, the anti-sense oligonucleotide ctelo was 5′-labelled with [γ- 32 P]ATP (6000 Ci/mmol, PerkinElmer) and T4 polynucleotide kinase (New England BioLabs) according to the manufacturer’s instructions. To generate dsDNA telomeric ligands, sense and anti-sense DNA strands at 2 μM concentration each were combined in annealing buffer (70 mM Tris-HCl (pH 7.6), 10 mM MgCl 2 ). Reactions were heated at 100°C for 5 min on a PCR thermocycler and subsequently cooled to 25°C at a constant rate of -1.0°C/min.

Results

Cryo-EM analysis Cryo-EM analyses of the purified and fully assembled shelterin complex (POT1, TPP1, TIN2, TRF2 and RAP1) (see Materials and Methods) were carried out on the unbound complex as well as the shelterin complex bound to a “model” telomere ligand (S1A Fig), containing both the appropriate double and single-stranded binding sites for the DNA-binding shelterin components POT1 and TRF2. However, despite evidence that the shelterin complex was fully active in DNA binding and telomerase activity (S2C Fig), all our original cryo-EM trials produced particles that appeared broken and did not cluster into 2D classes. While we were able to improve particle quality through mild glutaraldehyde crosslinking using the GraFix method [43, 44], 2D classification from large datasets produced only poor 2D classes that did not reconstruct to reliable 3D maps (S4 Fig). Additional efforts in sample and grid optimization did not improve particle homogeneity. Thus, we concluded that the high conformational heterogeneity of the full shelterin complex would prevent structure determination. On the other hand, cryo-EM analysis of the POT1-TPP1-TIN2(1–354) sub-complex appeared more promising, as our trials using first negative stain and subsequently cryo-EM produced homogeneous monodisperse particles that yielded good 2D classes. However, obtaining cryo-EM data that would allow 3D reconstruction was problematic. Firstly, the particles appeared smaller in size than expected given the molecular weight of the complex (169 kDa) and secondly, they suffered from severe orientation bias impeding 3D reconstructions. Fixation of the POT1-TPP1-TIN2(1–354) complex with crosslinkers or addition of telomeric DNA ligands, would consistently result in substantial degradation of the particle quality, a perplexing effect that could not be countered despite extensive optimization efforts. We hypothesized that the absence of DNA could be exposing a site on the protein complex that promotes sequestration of the particles to the air-water interface. Our hypothesis was supported during cryo-EM screening by the observation that very low concentrations (~0.05 mg/ml) of the POT1-TPP1-TIN2(1–354) sample used to prepare holey EM grids was consistently resulting in very high particle density. Therefore, to counter the effects of a possible adsorption of the particles to the air-water interface, we prepared affinity grids using graphene layered holey grids soaked with pyrene-PEK1K-(Ni)NTA to capture uncleaved POT1-TPP1-TIN2(1–354) complex through its His-10 tag. These grids offered a novel composition of heterogeneous particles of various shapes and sizes (10–20 nm), but did not produce satisfactory 2D classes. We therefore concluded that the otherwise undesirable effect causing our particles to potentially cluster to the air-water interface was actually a useful one, since only under these conditions were we able to obtain satisfactory 2D classes. Thus, we concentrated our efforts on holey UltrAuFoil grids that produced monodisperse POT1-TPP1-TIN2(1–354) particles, but with severe preferred orientation of the complex (Fig 2A). To overcome the orientation bias and to obtain better 3D information we collected data at a 0° and 30° tilt angle of the electron microscope stage. We also used the Volta Phase Plate (VPP) [45] to improve contrast, because in our initial datasets collected with a tilted stage, the quality of the 2D classes was compromised, most likely due to the increase in ice thickness when imaging a tilted grid. Tilted and un-tilted data sets were combined, and micrographs manually selected for preprocessing. From these micrographs, we extracted 2,103,136 particles for 2D classification (Fig 2B), from which, 620,136 particles were chosen for further 3D classification (Fig 2C). After 3D classification, two predominant and similar classes were present, that represented 45.4% and 34.8% of the total number of particles. These two classes appeared to differ slightly in their shape and were therefore processed and refined independently resulting in the reconstruction of one map (EMD-30596) at 7.9 Å, and the other (EMD-30597) at 9.6 Å resolution. These two maps appeared to correspond to the same particle, but differed in shape. Furthermore, the volume of the maps was not sufficient to accommodate the full POT1-TPP1-TIN2(1–354) complex and hence the two cryo-EM maps represented only part of the complex. In this case, major structural heterogeneity may also be the main cause of the incomplete density maps. Nevertheless, the resulting 3D reconstructions support the existence of (at least) two alternative conformational states of the POT1-TPP1-TIN2(1–354) complex.

Two conformations for the DNA-binding domain of POT1 Because the available structural information on POT1 is limited to the X-ray structures of the two halves [DNA-binding domain (OB1-OB2) and TPP1-binding domain (OB3/HJR)] [23–25] of POT1, the model derived from our cryo-EM analysis provides the first view of the structure of full-length POT1. In full-length POT1, the N-terminal OB1-OB2 half pivots from the OB3/HJR domain at a ~45° angle (Fig 3B and 3C). Whilst the crystal structure of OB3/HJR bound by the PBD of TPP1 fits well into the cryo-EM map (Fig 3), the improved fit of the OB1 and OB2 domains when fitted separately suggests that full-length POT1 can adopt different conformations, particularly in the relative orientation of the OB folds when not bound to DNA (Fig 3B). OB1 and OB2 are connected by a short linker (S145-T149), and in the crystal structure of POT1 bound to DNA, the middle three amino acids (P146-W148) were unresolved, suggesting some flexibility. In the cryo-EM map, the unresolved density connecting OB1 to OB2 (Fig 4B) also suggests additional flexibility in that region. Thus, this flexible linker could potentially allow OB1 and OB2 to take up a tighter conformation when bound to DNA. The second reconstruction obtained from the same dataset for the POT1-TPP1-TIN2(1–354) complex provides additional insights. The resolution of this map is 9.6 Å, and therefore more difficult to interpret de novo. However, by making use of the fitting of structures into the 7.9 Å map, we were able to place each of the POT1 domains (OB1, OB2, and OB3/HJR) into the 9.6 Å map in similar relative orientations (Fig 5A). While it is impossible to discern some of the details that were apparent in the 7.9 Å map, what is clearly evident is that whilst most of the overall density of the 9.6 Å map overlays well with the 7.9 Å map, the relative orientations of the OB1 and OB2 folds are distinctively different to what was observed in the 7.9 Å map, or in the crystal structure of the POT1 OB1 and OB2 domains bound to the ssDNA ligand (S5B and S5C Fig). While in the 7.9 Å map, OB1 and OB2 take up a compact or closed conformation that resembles that seen in the crystal structure of the OB folds bound to DNA (closed conformation, Fig 5C), in the 9.6 Å map the density for OB1 is displaced relative to OB2 taking up an extended conformation (open conformation, Figs 5C and S5C). The global movement, measured as the centre-to-centre distance between the two OB folds (as defined from fitting the two globular domains into the two EM maps), increases by about 5 Å in the open state (Fig 5C). Furthermore, the relative orientation of the OB2 domain as it extends out of the OB3/HJR domain also appears to differ in the two maps, suggesting that the OB2 domain may also swing. This observation suggests that POT1 has a dynamic structure in which the three domains (OB1, OB2, and OB3/HJR) are flexibly linked (Fig 5B and 5C) resulting in alternative conformations of full length POT1 (S5 Fig). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. Cryo-EM map at 9.6 Å resolution of POT1 with fitted crystal structures indicates an alternative open conformation. (A) Simulated 8 Å maps generated with Chimera for POT1 OB1 (magenta) and OB2 (green) fitted into the 9.6 Å cryo-EM map suggest the two domains exist in an open/extended conformation. (B) Cartoon representation of full-length POT1 (yellow) bound by the PBD of TPP1 (blue) fitted into the 9.6 Å cryo-EM map. (C) Comparison between closed and open conformations of full-length POT1 complexed to the PBD of TPP1 observed through cryo-EM. The distance between the centres of OB1 and OB2 is increased by about 5 Å in the open conformation compared to the closed conformation. https://doi.org/10.1371/journal.pone.0264073.g005 In summary, we have been able to fit most of the two cryo-EM density maps that dominate the family of reconstructions for the POT1-TPP1-TIN2(1–354) complex. The fitting of the known structures of sub-domains into the 7.9 Å map demonstrates that the reconstruction we have obtained represents full-length POT1 bound by the PBD of TPP1. Comparison to the 9.6 Å map reveals a different conformation for the DBD of POT1 in which the two OB folds have moved apart and are in a more open conformation than the closed one observed in our 7.9 Å map. The alternative conformations of OB1 relative to OB2 indicates a flexibility in the DBD of POT1 that likely has implications for how POT1 binds to telomeric DNA repeats.

[END]

[1] Url: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0264073

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/