Tom1p is a phosphosphorylation-dependent binding partner of Spt6tSH2

A) Domain structures of Tom1p, Spt6p, and Rpb1p with 3D structures of the core RNAPII elongation complex (6GMH, Vos et al. 2018) and the tSH2:Rpb1p linker complex (5VKO; Sdano et al. 2017). The two phosphorylation-dependent binding pockets on Spt6ptSH2 are shown with phosphates colored red. Arrows indicate interactions of Tom1p, Spt6p and Rpb1p. Domains of Tom1p described in subsequent figures are named.

B) Yeast strain 7382-3-4 was transformed with vector (Streptavidin-FLAG alone) or plasmid pMS14 (Streptavidin-FLAG-Spt6ptSH2), and whole cell lysates were subjected to tandem affinity purification. Proteins that copurified with the Spt6-tSH2 domain were visualized by Coomassie Blue stained SDS-PAGE (Protein) or by detection with purified GST or GST-Spt6ptSH2 and anti-GST antibodies (Far-western) after transfer to nitrocellulose. Mass spectroscopy was performed on bands excised from the gel after treatment with trypsin, revealing the identities of the proteins indicated.

C) Purified Tom1p (Tom1p-Protein A after Protein A removal) was treated with phosphatase (CIP) as indicated. Parallel SDS-PAGE replicas were stained with Coomassie Blue (Protein) or transferred to nitrocellulose and probed with Spt6ptSH2 (Far-western). Three repeats were performed with three separate samples, and binding was quantified with each sample normalized to the value obtained without CIP. Overall, phosphatase reduced binding to 18% of the untreated values (P = 4 x 10-6).

D) As in panel C, purified Tom1p was subjected to SDS-PAGE with or without phosphatase (CIP) treatment and probed with WT GST-Spt6ptSH2 or variants with mutations in one or both of the binding pockets, see panel A: -R indicates an R1282H mutation in the …TPY… binding pocket, -KK indicates a double K1355A, K1435A mutation in the …FSP… binding pocket, and -R,KK indicates all three mutations as described previously (Sdano et al. 2017). Results from multiple independent measurements were normalized to the average value for WT Spt6-tSH2 with the average and standard deviation shown (t tests yielded P < 0.01 = **, < 0.001 = ***, < 0.0001 = ****).

Spt6ptSH2 binding maps to the acidic domain of Tom1p

A) Purified Tom1p was incubated with trypsin. Aliquots removed at the indicated times were visualized by SDS-PAGE followed by Coomassie Blue staining (left) or far-western probing with GST-Spt6ptSH2.

B) The 60 kD band indicated with an arrow in (A) was excised and subjected to complete digestion with trypsin followed by mass spectrometry. The locations of peptides identified are shown in panel B. The inferred tSH2 binding site (bracket) starts with the first peptide of the most significant cluster of peptides and extends about 60 kDa, including the acidic domain which contains few cleavage sites for trypsin.

C) Markerless in-frame deletions were constructed to remove the entire acidic domain (intΔAD) or subregions within this domain at the endogenous TOM1 locus. Whole cell lysates were prepared from log-phase cultures and subjected to SDS-PAGE followed by detection with antibodies against Tom1p (middle panel) or Spt6ptSH2-GST fusion protein (bottom panel). The antisera cross-reacted with an unidentified protein of about 250 kDa (asterisk) which was unrelated to Tom1p as it was retained in a strain lacking the entire TOM1 gene (tom1-Δ0).

The Tom1p acidic domain is important for in vivo functions

Yeast strains with the mutations shown in Figure 2C were grown to saturation, washed in water, and 10-fold serial dilutions were placed on the media shown and incubated as indicated. YPAD is yeast extract and peptone supplemented with adenine and 2% dextrose and, as indicated, with 15 mM caffeine, 3% formamide, or 6 µg/mL phleomycin. Gal indicates substitution of the dextrose with galactose.

The acidic domain of Tom1p mediates binding to histones and nucleosomes

Versions of Tom1p fused to Protein A were purified as described in Methods, mixed with purified yeast histones alone or after reconstitution into nucleosomes with a 181 bp DNA fragment, then separated by native PAGE (Xin et al. 2009). Histones and nucleosomes were detected by a fluorescent dye attached to a unique cysteine residue in H3 or H2B, or at the 5’ end of the DNA. Free histones do not migrate uniformly in this gel system, typically forming aggregates in the well or being lost during analysis, so bands indicate the formation of complexes with Tom1p. Nucleosomes retain variable amounts of H2A-H2B dimer, with octamers (nucleosomes) and hexamers (hexasomes) migrating differentially, as indicated. Similar to the results with FACT and Spt6p (Xin et al. 2009, McCullough et al. 2015), nucleosomes only formed complexes with Tom1p in the presence of the HMGB family factor Nhp6p.

Genomic analysis supports a role for Tom1p in transcription that involves maintaining chromatin architecture in promoters and gene bodies

A) RNA-seq was performed with WT and matched tom1-Δ0 strains in triplicate and the log2FC value for mutant/WT was calculated for each gene. Transcript levels from most genes were only modestly affected by loss of Tom1p (7 genes had decreased transcript levels and 16 genes were increased at a threshold of a 2-fold decrease at a false discovery rate below 1%). As a group, transcripts from genes encoding ribosomal protein genes (red stars) were generally decreased, forming a distinct subset (red stars).

B) A 12xV5 tag was fused to the C-terminus of the Tom1p ORF at the endogenous locus and anti-V5 antibodies were used to perform ChIP-seq with three independent clones and a matched strain without the tag. ChIP signal relative to input DNA was calculated genome-wide and mapped for each gene normalized to an averaged gene size of 1000 arbitrary units, as well as 200 bp upstream and downstream of each gene. Genes were ordered by their transcript levels determined by RNA-seq using a matched WT strain (Connell et al. 2022 and panel A). The average Tom1p occupancy values are shown for the highest, lowest, and middle transcript level deciles.

C, D) MNase-seq was performed in triplicate with independent isolates of WT and tom1-Δ0 strains, and reads were parsed into groups by the size of the protected DNA fragment. 140-170 bp fragments were classified as nucleosomes, representing ∼36% of all reads, and 90-120 bp fragments were classified as subnucleosomes, representing ∼3.6% of all reads. Protected fragments were mapped to the annotated transcription start site (TSS) of each gene. Genes were placed in deciles by transcript levels as in panel B and the log2FC occupancy values shown are relative to nucleosomes (C) and subnucleosomes (D).

Structure of yeast Tom1p in the closed conformation and comparison with human HUWE1

A) Alignment of human HUWE1 and Tom1p sequences. Modeled residues indicated with solid bars; disordered regions are striped. Motifs/domains recognized from the sequence or defined in the structure are indicated with color. Tom1p: Head domain (1-436), Neck (437-506), UBA-like domain (1195-1261), HWA(1510-1594), acidic domain (1873-2131), Tower helix (2221-2276), UBM module (2317-2398), HECT N-terminal lobe (2889-3146), HECT C-terminal lobe (3147-3268). Body colored grey. Human HUWE1 WWE domain is not present in Tom1p. Beginning and end points of missing density in Tom1pΔAD structure: 214-235, 1873-2219 (region containing acidic domain), 2276-2316, 2399-2449, 2575-2589.

B) Top-down (left) and side (right) views of the 3.07 Å Tom1pΔAD map. Disordered residues of the acid domain are shown as pink coil in random positions. Other disordered regions not depicted. The structure of Tom1pHis is very similar, with the exception of more poorly resolved density at the UBM domain.

C) Top-down view of Tom1pΔAD (grey) aligned with human HUWE1 (tan) (Hunkeler et al. 2001; PDB 7JQ9). The close-up shows a clipped side view of the interior of the basket between the Tower and UBM regions. The ordered residues either side of the disordered acidic domain are shown: Tom1p, magenta, residues 1872, 2220; HUWE1 light pink).

Tom1pHis hinging motion

A) When aligned separately, the N-terminal and C-terminal halves of closed, intermediate, and open conformation Tom1 superimpose closely. Center: Top-down view of the Tom1pHis closed-conformation model with residues 1-1315 colored dark grey and 1321-3268 colored light gray. Top: same orientation showing the three Tom1pHis conformation models aligned for residues 1-1315: closed (grey), intermediate, (pink) and opened (turquoise). Bottom: Same orientation, showing the three Tom1pHis conformation models aligned for residues 1321-3268.

B) The N-terminal and C-terminal halves of closed, intermediate, and open conformation Tom1p display a hinging motion Top-down and side views of the three Tom1pHis conformation models aligned on residues 1-1315. The hinging motion, centered on loop 1316-1320 (green), is ∽ 15° and ∽30° for the intermediate and open conformations, respectively.

C) Same as (B) but showing maps for the three structures.

HECT domain comparisons between constructs, conformations, homologs, and catalytic states.

A) Tom1 HECT domain. Tom1pΔAD HECT domain is shown. Tom1pHis closed, intermediate, and open conformations are superimposable within experimental error. Catalytic cysteine, yellow.

B) Superposition of Tom1p and HUWE1 HECT domains. Tom1pΔAD, gray; human HUWE1, tan (Hunkeler et al. 2021; PDB 7MWD), Tom1p HECT crystal structure, green.

C) Tom1p (grey) and human Nedd4L HECT domains superimposed on their N-lobes. The Nedd4L HECT structure (lavender) is a complex with a Ub-charged E2 (yellow, green) (PDB 3JW0; Kamadurai et al. 2009. C-lobes show a relative rotation of 63°.

D) Tom1pΔAD (grey) superimposed on N-lobe with human HUWE1 HECT domain (brown) covalently bound to Ub-propargylamine (yellow) (PDB 6XZ1; Nair et al. 2021). C-lobes show a relative rotation of 167°.