Structural basis of Nrd1-Nab3 heterodimerization

Abstract

Heterodimerization of RNA binding proteins Nrd1 and Nab3 is essential to communicate the RNA recognition in the nascent transcript with the Nrd1 recognition of the Ser₅-phosphorylated Rbp1 C-terminal domain in RNA polymerase II. The structure of a Nrd1-Nab3 chimera reveals the basis of heterodimerization, filling a missing gap in knowledge of this system. The free form of the Nrd1 interaction domain of Nab3 (NRID) forms a multi-state three-helix bundle that is clamped in a single conformation upon complex formation with the Nab3 interaction domain of Nrd1 (NAID). The latter domain forms two long helices that wrap around NRID, resulting in an extensive protein-protein interface that would explain the highly favorable free energy of heterodimerization. Mutagenesis of some conserved hydrophobic residues involved in the heterodimerization leads to temperature-sensitive phenotypes, revealing the importance of this interaction in yeast cell fitness. The Nrd1-Nab3 structure resembles the previously reported Rna14/Rna15 heterodimer structure, which is part of the poly(A)-dependent termination pathway, suggesting that both machineries use similar structural solutions despite they share little sequence homology and are potentially evolutionary divergent.

Figure 1.. Structural data for the isolated Nrd1 and Nab3 heterodimerization domains.

(A) Schematic representation of Nrd1 and Nab3 domain architecture with the heterodimerization domains NAID and NRID highlighted in blue and red, respectively. Sequence logos to represent the amino acid conservation of these domains have been produced from sequence alignments of Nrd1 and Nab3 orthologs of organisms of the Saccharomyces clade (full sequence alignments in Fig S1). Other domains/regions are displayed: CID (CTD interaction domain), RBD (RNA binding domain), ABD (tRNA anticodon binding domain), DE-rich (acidic region), and RE-rich (arginine/glutamic-rich region). (B) ¹H-¹⁵N HSQC spectra of Nrd1 NAID (residues 147–222) (left panel in blue) and Nab3 NRID (residues 191–261) (right panel in red) in their isolated forms. (C) Percentage of secondary structure calculated from ¹³C/¹H chemical shifts for Nrd1 NAID (left panel) and Nab3 NRID (right panel). The bar charts indicate the percentage of α-helix (blue/red bars) versus random coil (grey bars) calculated with the program d2D+ (Camilloni et al, 2012). Other types of secondary structures have been omitted because of their low calculated percentages. Nrd1 NAID residues with missing HQSC cross-peaks are indicated with stars. (D) Superposition of the circular dichroism spectra of Nrd1_147-222 (in blue) and Nab3_191-261 (in red). Black arrows mark the position of the two typical minima at 208 and 222 nm exhibited by α-helix structures. (E) Superpositions of the 20 lowest target function conformers calculated for Nab3 NRID (residues 198–250) (PDB code: 7PRE). Structures have been optimally superimposed considering only the N-terminal α-helix (residues 208–221) (right panel) or the C-terminal α-helix (residues 239–246) (left panel). The relative orientation of the two α-helices is loose and only minimally constrained by the interactions between side chains of residues Val₂₁₅, Ile₂₄₁, and Phe₂₂₉ (labelled and colored in green, hydrophobics, and pink, aromatic).

Figure S1.. Comparison between amino-acid sequences of heterodimerization domains of Nrd1 and Nab3 orthologs of Saccharomyces cerevisiae and close-related fungi.

(A, B) Sequence alignments of Nrd1 NAID (A) and Nab3 NRID (B) for yeast species to Saccharomyces clade: Tetrapisispora phaffii (TETPH), Tetrapisispora blattae (TETBL), Vanderwaltozyma polyspora (VANPO), Naumovozyma castellii (NAUCC), Naumovozyma dairenensis (NAUDC), Candida glabrata (CANGA), Torulaspora delbrueckii (TORDC), Zygosaccharomyces rouxii (ZYGRC), Kazachstania africana (KAZAF), Lachancea thermotolerans (LACTC), Ashbya gossypii (ASHGO), and Kluyveromyces lactis (KLULA). Protein database codes are as follows: Nrd1_YEAST (P53617), Nrd1_TETPH (G8BQ11), Nrd1_VANPO (A7TF47), Nrd1_ NAUCC (G0V5A0), Nrd1_NAUDC (G0WD58), Nrd1_CANGA (Q6FNZ7), Nrd1_TORDC (G8ZSI9), Nrd1_ZYGRC (C5DYV5), Nab3_YEAST (P38996), Nab3_TETBL (I2GYZ3), Nab3_TETPH (G8C176), Nab3_VANPO (A7TJ31), Nab3_KAZAF (H2B198), Nab3_NAUCC (G0VIS8), Nab3_CANGA (Q6FS59), Nab3_TORDC (G8ZYP4), Nab3_ZYGRC (C5DZZ2), Nab3_LACTC (C5E2G1), Nab3_ASHGO (Q754Y1), and Nab3_KLULA (Q6CML8).

Figure S2.. Detailed view of the 2D NOESY of Nab3 NRID (residues 191–261) (in 100% D₂O) showing the NOE cross-peaks of the aromatic protons of residue Phe₂₂₉ with methyl groups.

Medium-range cross-peaks with Leu₂₂₄ and Leu₂₂₆ typical of α-helix structures are shown, together with long-range ones with Ile₂₄₁ and Ala₂₄₀.

Figure 2.. Nuclear magnetic resonance (NMR) and thermodynamic analysis of Nrd1–Nab3 heterodimerization.

(A) Superposition of the ¹H-¹⁵N-HSQC spectra of Nrd1 NAID (residues 147–222, left panel) in its free form (grey) and after addition of unlabelled Nab3 NRID (residues 191–261) (blue). Analogous NMR spectra comparison for ¹⁵N-labelled Nab3 NRID (right panel) showing the superposition of free (grey) and Nrd1 NAID-bound (red) NMR spectra. Unlabelled proteins were added in excess to ensure the saturation of the labelled ones. (B) Bar charts showing the per-residue population percentage of α-helix (blue/red) and random coil (grey bars) secondary structure for bound forms of Nrd1 NAID (left panel) and Nab3 NRID (right panel). (C) Isothermal titration calorimetry analysis of two different Nrd1–Nab3 interactions. The Nab3_191-261 construct was titrated over two Nrd1 constructs (left: Nrd1_1–222 and right: Nrd1_{147-222/290-489}) including the domains shown in the scheme. Thermograms (upper panels) and binding isotherms (lower panels) are shown for each titration, together with the equilibrium dissociation constant K_D(1/K_B), enthalpic (ΔH), and entropic contributions (ΔS), and stoichiometry (N) values calculated from data fitting to one-site binding model. Experiments were performed at 15°C.

Figure S3.. Additional ITC experiments of Nrd1/Nab3 hetererodimerization.

(A) Replicas of the isothermal titration calorimetry experiments shown in Fig 2C. (B) Isothermal titration calorimetry experiments obtained by titrating Nab3 NRID (residues 191–261) over the protein construct txA-HTEV-Nrd1_147-222 NAID. The N-terminal fusion protein contains the sequence of E. coli thioredoxin A followed by a 6xHis tag and a consensus TEV cleavage site.

Figure S4.. NMR data comparison accross diferent Nrd1-Nab3 chimeras.

(A) Superposition of the ¹H-¹⁵N HSQC spectra of the Nrd1_147-222-Nab3_202-261 chimera (grey peaks), Nrd1 NAID (blue) and Nab3 NRID (red) on their bound forms. The peaks showing the largest differences are marked and labelled. (B) Backbone amide chemical shift differences (CSD = (Δδ_NH² + Δδ_N²/5)^1/2) between Nrd1(blue bars)/Nab3(red bars) peaks in the heterodimer and in the different chimeric constructs tested during the optimization process. Outlines of the different constructs are represented above the graphs. Dashed lines (zero length) and grey box show connecting linkers between the two parts of chimeras.

Figure 3.. Nuclear magnetic resonance structure of the Nrd1–Nab3 chimera.

(A) ¹H-¹⁵N HSQC spectrum of the Nrd1_158-222-Nab3_202-261 chimera recorded at 800 MHz and 25°C. Cross-peaks assignments have been labelled according to the amino acid sequences of Nrd1 (in pink) and Nab3 (in cyan) fragments that compose the chimera. The horizontal lines mark the two cross-peaks of amide NH₂ moieties in side chains of Gln and Asn residues. (B) Superposition of the 20 structural models calculated by nuclear magnetic resonance (statistics in Table S1) (upper panels) (PDB code: 7PRD). Two different orientations are shown. The Nrd1 in the chimera is colored in light pink and the Nab3 part in light cyan. Regular secondary structure elements are named consecutively (α-helices α1 to α5). Surface representations of the structure in the two selected orientations and with the same color code are shown below. (C) Structural details of the interaction between Nrd1 and Nab3 parts of the chimera. Only side chains of residues involved in heterodimeric contacts are shown. The interface is mainly formed by hydrophobic residues with the exception of Nrd1 Gln₂₀₅ and Gln₂₁₇ with Nab3 Gln₂₁₄ and Asn₂₂₅ that participate in two hydrogen bond networks (yellow dashed lines) that are buried inside the structure. The Nab3 Phe₂₂₉-Ile₂₄₁ contact, present in the free form (Fig 1E), is maintained in the Nrd1–Nab3 chimera. (B) Residues have been numbered according to the Nrd1 and Nab3 sequences and colored as in panel (B).

Figure S5.. Nuclear magnetic resonance data obtained with selective ¹³C methyl labelling of Leu, Val, and Ile.

(A) ¹H-¹³C HSQC spectrum with assignments. Methyl cross-peaks assignments have been color coded according to the protein segment they belong in the chimera: Nrd1 (in cyan) and Nab3 (in pink). (B) Selected ¹³C–¹³C planes of the 3D ¹H-¹³C-HSQC-NOESY-¹H-¹³C-HSQC spectrum showing various methyl–methyl NOEs.

Figure S6.. Interactions between Nab3 Tyr₂₁₇, Ser₂₄₇ and Nrd1 Arg₁₇₃.

(A). Selected region of the 2D NOESY showing the resonance Tyr₂₁₇ Hη and NOEs with other Tyr₂₁₇ ring protons and with side chain resonances of Arg₁₇₃. (Β). Detailed view of the superposition of the 20-conformers of the Nrd1–Nab3 chimera, showing the side chains of Nab3 Tyr₂₁₇, Ser₂₄₇, and Nrd1 Arg₁₇₃.

Figure 4.. Functional analysis of Nrd1/Nab3 heterodimerization.

(A) Scheme representing the distribution of the analyzed mutants (indicated as green starts). Six positions in Nrd1 NAID domain were mutagenized (see specific details in the text). (B) The six mutagenized residues in Nrd1 NAID correspond to hydrophobic amino acids (Leu₁₈₉, Leu₁₉₃, Leu₁₉₇, Leu₂₀₉, Ile₂₁₃, and Leu₂₁₆) buried in the structure. These Leu or Ile side chains were replaced with Ala (conservative mutation) or Arg (disrupting mutation). (C, D). Analysis of the growth phenotypes of the nrd1 mutants and wild-type cells (wt.). The temperature-sensitive mutant nrd1-K335E, previously identified in the RNA-binding domain (Franco-Echevarría et al, 2017), is included as reference. Cultures were serially diluted (1/10), spotted on selective SC media plates and grown at the indicated temperatures for 2–3 d. (C) The first set of mutants (Leu/Ile to Ala) does not show differential behavior compared to wt. at the two tested temperatures. In comparison, the nrd1-K335E temperature-sensitive mutant shows the expected growth phenotype at 37°C (Franco-Echevarría et al, 2017). (D) Among the second set of mutants, including Leu/Ile to Arg mutations, nrd1-L209R and nrd1-I213R show strong growth defects, even lethality at 34°C and 37°C for nrd1-L209R mutant. Two clones of each mutant were tested.

Figure 5.. Structural comparison between CFI and NNS complexes.

(A) The structural models depict the current knowledge about the organization and interactions within the Cleavage Factor I and Nab3–Nrd1–Sen1 complexes, that are involved in the two transcription termination pathways in yeast (see the Introduction section for details). On the right, termination of short transcripts is associated to Ser₅ phosphorylation mark in RNA Pol II (blue dots in the schematic representation of Rpb1 CTD) that are recognized by Nrd1 CID (PDB: 2IO6 in pink and Rpb1 CTD in grey/blue [Ser₅-P]). On the nascent transcript, Nrd1 (PDB: 5O1Y in pink) and Nab3 (PDB: 2L41 in cyan) RNA-binding domains recognize specific terminator sequences (black line and boxed RNA sequences below). The helicase Sen1 (PDB: 5MZN) also recognizes unspecific RNA sequences, and its intrinsically disordered region contains three Nrd1 interaction motifs (NIMs): NIM1, NIM2, and NIM3 (marked in red) that can interact with the CID, competing out the Rpb1 CTD and allowing the termination process to evolve to its final steps (Zhang et al, 2019; Han et al, 2020). On the left, CFI uses similar strategies. The CID of Pcf11 (PDB: 1SZA in purple) recognizes Ser₂-P CTD-derived peptides (yellow dots and yellow atoms in the 1SZA structure), typical of long-elongated transcripts, whereas Hrp1 (orange) and Rna15 (maroon) (PDB: 2KM8) recognize the polyadenylation signal and enhancement elements. Clp1 (grey) recognizes a Pcf11 peptide (in purple) (PDB: 2NPI) and also interacts with other proteins of CFI (yet-unknown structures). The Rna14 HAT domains (yellow) interact with Hrp1 RRMs (Barnwal et al, 2012) and its Monkeytail domain forms a heterodimer with the C terminus or Rna15 (maroon) (PDB: 2L9B). This heterodimer has a similar structure as the Nrd1–Nab3 chimera (PDB: 7PRD this work). (B) Comparison between the structures of Rna14/Rna15 heterodimer and Nrd1–Nab3 chimera. In both cases, models have been represented as a surface/ribbon mixture for each of the components, and alternating between them in top and bottom figures (identical orientation for each structure). Rna14 Monkeytail domain (yellow) and Nab3 interacting domain in Nrd1 (pink) wrap around their partners in a similar way, creating large protein–protein interfaces. In the structures, Rna15 (maroon) and Nab3 (cyan) form compact helix bundles.

Figure S7.. Evolutionary reconstruction of the Nrd1 NAID.

The sequences of 121 Nrd1 orthologues were obtained from (https://omabrowser.org/) (Altenhoff et al, 2020) and aligned using the full-length proteins. Higher levels of conservation are found on RNA-binding domain and CID domains. A subregion comprising the Saccharomyces cerevisiae Nrd1 NAID (residues 161–220) of the alignment was extracted and ranked according the phylogenetic tree on the left. The tree was obtained with http://www.timetree.org (Kumar et al, 2017). Only 45 of the original 121 Nrd1-like proteins (codes next to the species name) are represented, corresponding to those species with match in the TimeTree database. The phylogenetic tree includes a geologic timescale with a time line and other various indicators. The position of Saccharomyces cerevisiae Nrd1 is highlighted in grey and the branches corresponding to the Saccharomyces and its close Candida clades are labelled in the tree. Below the alignment, the structural elements of Nrd1 NAID have been colored in green (helix α1), cyan (extended segment contacting helix α2), and red (helix α2), with the Nab3 NRID representing the surface. The boundaries of these elements have been shadowed with the same color code over the alignment above.