Ribosome SUMO Viewer

What this shows

The human ribosome with the SUMO2/3 acceptor lysines of its proteins, on a choice of structures: the mature 80S, translating and idle ribosomes, two collided ribosomes, and assembly intermediates of both subunits. On the structure shown now (, PDB ) sites on ribosomal proteins are listed; of them have coordinates and are drawn as enlarged, coloured residues. Each site carries its response to proteasome inhibition (MG132) and to heat shock, and a description of where the lysine sits in the particle.

Eight steps

1. Choose a structure under Structure; the page opens on the mature 80S. Structures.
2. Look around. Drag to rotate, scroll to zoom. All controls.
3. Choose what the colours mean under Colour the sites by. The default is the MG132 fold change. Colourings.
4. Jump to a standard view — the 40S or 60S solvent side, the two faces between the subunits, top, bottom. Views.
5. Find a site. Type a protein name above the table, sort by a column, click a row: the structure centres on that lysine and its details appear below the table. Table.
6. Ask whether a SUMO fits. The colouring Room for SUMO shows at which sites a SUMO has space; the details of a site can draw one there. Room for SUMO.
7. Share what you see. Copy link gives an address that opens the same structure, colouring, filters, site and view. Sharing a view.
8. Add your own numbers. Load a table… colours proteins or sites with values from a file on your computer; the file stays on your computer. Your data.

Reading the picture

Every atom is drawn as a sphere. rRNA is light grey, the proteins of the small (40S) subunit sand, those of the large (60S) subunit blue-grey. Factors bound to the ribosome are olive, tRNA teal, mRNA black. SUMO sites are drawn larger and in colour, so that a single lysine remains visible when the whole ribosome is in view.

The enlargement is a marker, not the size of the residue. A site that is buried inside the particle shows only where its marker reaches the surface, or not at all; the table lists it all the same, and selecting it moves the view to its position.

Sharing a view

The address of the page holds what is shown. Whatever you choose is written into it as you go, so copying the address, or pressing Copy link at the top, gives a link that opens the viewer in the same state on another computer.

What is at its default is left out, so the link of the opening view is the bare address. A camera move is written a moment after it has come to rest.

Not in the link: a table of your own, and the colouring of proteins by it. The table is a file on your computer and is never part of an address; send the file along and load it on the other side. Changes made in Mol*'s own panels are not in the link either.

Keyboard

? opens this help, Esc closes it. Every ⓘ on the page opens the help at the entry for the control next to it. The address of an open entry (for example #help/env/classes) can be copied and shared, like the rest of the address: see Sharing a view.

Choosing a structure

A structure is one state of the ribosome, and a lysine that is enclosed in one state can be free in another. The Structure menu therefore offers several, all of the human ribosome and all from the PDB:

Choosing a structure replaces the picture, the site table and the Statistics window. The colouring, the selected site and a table of your own carry over. The address then holds s= and the PDB code (for example ?s=8G60), beside everything else that is chosen, and can be shared (Sharing a view); without it the page opens on the mature 80S.

The mature 80S is the reference: it has the highest resolution, and the statistics on regions, clustering and hotspots are computed on it. The decoding 80S adds mRNA, two tRNAs and eEF1A, and resolves the L1 stalk (RPL10A), RPL10 and part of the P stalk base (RPL12), which the mature structure does not. The idle 80S carries eEF2 and SERBP1, the state of ribosomes that are not translating. The collided disome is a stalled ribosome with the next ribosome run into it, bound by the kinase ZAK and by EDF1. The pre-60S and pre-40S particles are subunits in the making, with assembly factors in place of what is still to come.

Any other human protein can be shown as well, as its predicted model from the AlphaFold DB with its sites from the same catalogue (Any protein from the AlphaFold DB), and so can any entry of the PDB, with the sites of all the proteins in it (Any entry of the PDB).

One site in several structures

Each structure lists the sites of the ribosomal proteins it contains. The numbers differ because the structures differ: a precursor of the large subunit has no 40S proteins, the decoding ribosome has stalk proteins that the mature structure lacks, and the disome holds every protein twice, so each of its sites appears once for the stalled and once for the collided ribosome.

The detail card of a selected site has a line Elsewhere: the environment of the same lysine in each of the other structures, or not modelled where the residue has no coordinates there, or protein absent. A name in that line opens that structure with the same site selected and in view. For the disome the line gives the better-resolved of the two copies.

The filter resolved here, not in the mature 80S lists what a structure adds to the reference.

Any protein from the AlphaFold DB

The box under the Structure menu takes any human protein — a gene name such as RANGAP1, or a UniProt accession such as P46060 — and shows the predicted model of that protein alone from the AlphaFold DB, with its SUMO2/3 sites from the same catalogue as the ribosome's (Hendriks et al. 2018: sites on human proteins). The name is looked up at UniProt (the reviewed human entry of that gene; other matches are offered under the box), the model is fetched from the AlphaFold DB, and the sites are placed on it as they are placed on the ribosome: a site goes to the lysine that carries its sequence context — at the catalogue's position where the model has it there, elsewhere when the model numbers the protein differently (the detail card says so), and not at all when the model has no lysine in that context (not placed in the table). Then the viewer computes the room for a SUMO at every lysine of the model, in your browser: see Room for SUMO computed in your browser.

Under the box, the note lists the experimental structures of the protein in the PDB (from UniProt: method, resolution, the chain and the residues it covers), sorted by resolution; a code opens that entry with the sites of all its proteins (Any entry of the PDB). A last line names the curated complexes the protein belongs to in the Complex Portal, which records composition, mostly without coordinates.

A protein offers the colourings by response (the fold changes and the Z-scores) and Room for SUMO on the model, the control SUMO at the sites, the focus, the legend switches, the confidence colouring in the banner, a table of your own, and the link (p= and the accession). What belongs to the ribosome structures is not there: the environment, region, hotspot and window colourings, the partners, the standard views and the Statistics window. For a ribosomal protein the detail card keeps the line On the ribosome, which opens each ribosome structure at the site; Back to the ribosome in the banner returns to the structure shown before.

A model is a prediction of the protein alone, over its whole sequence, with the tails that no structure resolves; The free protein says how to read room on it. The AlphaFold DB has models for proteins of up to about 2,700 residues; a longer protein, or one the database has no model of, cannot be shown. Nothing is uploaded: the lookups go to UniProt and the AlphaFold DB, and the calculation runs in your browser.

Any entry of the PDB

A PDB code in the box under the Structure menu (3UIO, say), or a code from a protein's list of structures, opens that entry: a protein in its experimental complexes, the counterpart of the predicted protein alone. The coordinates come from the PDB as for the ribosome structures; the molecules of the entry and the UniProt protein each chain is a piece of come from PDBe (SIFTS); the sites come from the catalogue and are placed on every chain that maps to a human protein, so a complex shows the SUMO2/3 sites of all its members, each copy of a protein separately (the chain is tagged in the table and the card). A site goes to the residue that carries its sequence context in the chain's sequence — at the mapped position where the context agrees, elsewhere when the construct is numbered differently, with one substitution accepted; the card says how. A site whose residue has no coordinates in the entry is listed as not modelled.

Then the room for a SUMO is computed in your browser (Room for SUMO computed in your browser), with the chains kept apart: for every lysine the share of placements that fit in the entry as it is and on the lysine's own chain alone, the other chains taken away. A site with no room in the entry and room on its chain alone has its room taken by the other chains: a SUMO there stands in the way of the complex, or has to be removed first — the same reading as What takes the room away on the ribosome. The colouring Room for SUMO paints the room in the entry, Room on the chain alone the other; the card gives both shares, and Show where a SUMO would sit without the other chains draws the placement that becomes possible. The legend lists the molecules of the entry as switches; the Region menu filters by chain. In an entry with more than about 400 lysines only the sites' lysines are computed.

Not available for an entry: the environment, region, hotspot and window colourings, the partners checkbox, the standard views and the Statistics window; for a ribosomal protein the card keeps On the ribosome. Chains without a UniProt mapping (peptides, non-human proteins) are drawn but carry no sites. The link holds the entry as e= and the code.

Factors, tRNA and mRNA

Everything in a structure that is not ribosomal protein or rRNA is a partner: translation and assembly factors (olive), tRNA (teal), mRNA (black). The checkbox under the Structure menu takes them out of the picture, which uncovers the sites beneath them; in the nucleolar pre-60S, factors cover much of the surface.

In the disome the collided ribosome is drawn in paler tones than the stalled one, so that the two can be told apart.

A site counts as near a partner when its side-chain nitrogen (NZ) is within Å of an atom of a factor, a tRNA, the mRNA or, in the disome, of a protein or rRNA of the other ribosome. The nearest partner and the distance are in the detail card (Partner), in the tooltip, and can be coloured and filtered. A site is covered by a partner when its NZ would be exposed on the ribosome alone and is enclosed once the partners are in place.

Orientation

Every single-particle structure is superposed on the mature 80S through the phosphorus atoms of the rRNA they share, and its view buttons use the directions of the mature 80S. The same button therefore shows the same face of the ribosome in every structure. In a precursor of one subunit, a view named after the other subunit looks at the face where that subunit will bind.

The disome has views of its own; see the standard views.

Comparing with the mature 80S

Two tools set a structure against the reference. The colouring Environment in the mature 80S paints each site with the environment the same lysine has in the mature ribosome; switching between it and Lysine environment shows what changes. The filter exposed here, enclosed in the mature 80S lists the sites that are exposed in the structure shown and buried or in contact with rRNA in the mature one.

For the assembly intermediates the Statistics window counts these cases and compares them with the other lysines: see Against the mature 80S.

Mouse and touch controls

The complete list of key and mouse bindings is under Settings / Controls Info in the toolbar at the right edge of the structure.

Focus and overview

The two buttons at the bottom right of the structure switch between a close-up and the whole. Focus moves the camera onto the marked residue — the one last clicked, or else the site selected in the table — to about 70 Å, and opens a peek hole: within a narrow cone around the line of sight, what lies in front of the residue fades — atoms just in front of it are translucent, atoms farther in front more so, and from about 13 Å in front they are not drawn at all. Nothing is cut: the residue, what is level with it, what is beside the cone and everything behind it stay solid, so the surroundings of the lysine are seen as they are. The fade follows the camera as you turn and zoom. Overview brings the whole structure back, seen from the same direction. The buttons are available as soon as a residue is marked.

Choosing a site in the table focuses on it at once. A double click in the structure only moves the camera closer, without the fade, and a view button returns to the overview. In the focus, rotating and zooming work as usual, and a click on another residue moves the focus there.

The legend as switches

Every entry of the colour legend under the colouring buttons is a switch: click it to hide what it stands for in the structure — the sites of one class (or, for a fold change or Z-score, the coloured sites or the grey ones), the rRNA, the proteins of one subunit, the paler second ribosome, the factors, tRNA or mRNA, the SUMO placed at a site, the free protein — and click it again to show them. A hidden entry is struck through. The site table is not affected; the link carries the hidden entries (hide).

The standard views

The view buttons turn the ribosome to fixed orientations, derived from the structure itself so that they mean the same for every colouring and, through superposition, for every structure:

The subunit axis runs through the centres of the two subunits; "up" is the 40S head (the position of RACK1) or, for the 60S view, the central protuberance (5S rRNA).

The collided disome has six views of its own, laid out along the line that joins the two ribosomes:

The toolbar at the right edge

These buttons belong to Mol*, the program that draws the structure. From the top: Reset Zoom brings the whole particle back into view; Screenshot saves the current picture as an image; Toggle Controls Panel opens Mol*'s own panels (sequence, components, measurements) for anything this page does not offer; Toggle Full Screen; Settings / Controls Info; Illumination.

Changes made in Mol*'s panels are replaced when a structure or a colouring is chosen or a table is loaded.

The selected site

Selecting a row in the table moves the camera to that lysine and outlines it in green; clicking a residue in the structure outlines it the same way and pins its label. The outline stays until another site or residue is selected, or until a click on empty space.

MG132 fold change

The change of SUMO2/3 modification of the site after proteasome inhibition with MG132, as log2 fold change against control (HEK293 cells; Hendriks et al. 2018). A value of 3 is an 8-fold increase.

Only sites with q < 0.05 are coloured, from pale sand (small change) through rose to dark indigo (large change). The scale runs from 0 to 8; larger values are shown in the colour for 8. Sites that do not change significantly are dark grey. of the sites change significantly, and every one of them increases.

Heat shock fold change

The same as the MG132 colouring for heat shock: the log2 fold change of the site's SUMO2/3 modification after heat shock against control (Hendriks et al. 2018), on the same scale from 0 to 8, for sites with q < 0.05 only; sites that do not change significantly are dark grey. of the sites change significantly under heat shock. A few significant sites decrease under heat shock; they take the colour of the low end of the scale, and the detail card gives their value.

Z-score in control, MG132 or heat shock

The Z-score places a site's SUMOylation in one condition relative to the same site in the other conditions: 0 is the site's mean across control, MG132 and heat shock, positive values (rose) are above it, negative values (indigo) below, in units of the site's standard deviation. Values are those of the published data set; the scale runs from −1.5 to 1.5.

Switching between the three Z-score colourings shows the same sites under each condition: nearly all are low in control and rise with MG132; a subset rises with heat shock. A Z-score compares conditions within one site. It does not compare one site with another; the fold change does.

Sites without quantification are dark grey.

Lysine environment

Colours each site by where its lysine sits in the structure shown: exposed, in contact with rRNA, or buried. The definitions are under Lysine environment.

Environment in the mature 80S

Offered for every structure other than the mature 80S. Each site is painted with the environment that the same lysine has in the mature ribosome; a site that has no coordinates there, or whose protein is not part of it, is dark grey. Compare with Lysine environment to see which lysines change: see Comparing with the mature 80S.

Room for SUMO

Colours each site by whether a SUMO fits at its lysine in the structure shown: room, tight or none; sites without a resolved side chain are dark grey. What the classes mean and how they are computed is under Room for SUMO.

When a SUMO fits

Colours each site by the class of its window along the ribosome's life — free protein only, precursor only, free subunit only, translation state, mature only, throughout, none found, flexible segment; see the classes.

Room on the free protein

Colours each site by whether a SUMO fits at its lysine on the AlphaFold DB model of the protein alone: room, tight or none; sites of proteins without a model are dark grey. Switching between this colouring and Room for SUMO shows which sites gain their room once the protein is out of the particle. See The free protein.

Near a partner

Offered when the structure holds partners. Sites within Å of a factor are wine red, of a tRNA green, of the mRNA indigo, and of the other ribosome of a disome rose; all other sites are dark grey. A site near several is given the colour of the nearest. The site colours differ from the colours of the partners themselves so that a site remains visible against the factor it touches. Definitions: Factors, tRNA and mRNA.

Region

Colours each site by the part of the ribosome it lies in: 40S head, 40S body, the central protuberance of the 60S, or the 60S body. How the regions are drawn is described under Regions.

Hotspots

Shows the groups of MG132-responsive sites that lie close together, each group in its own colour; all other sites are dark grey. Hotspots, heat shock does the same for the heat-shock-responsive sites (groups HS1, HS2, …), with the same rule and the same colours by rank. A hotspot is a description of where the modification is dense, not a tested unit: see Hotspots. The hotspots are defined on the mature 80S and keep their number and colour in every structure.

Other modifications

Marks the sites whose lysine UniProt also records as carrying another modification: ubiquitin, an acetyl or a succinyl group. A site with several is shown in the colour of the first of these that applies; the detail card lists all of them. Sites without such a record are dark grey. See Other modifications.

q-values and significance

q is the false-discovery-rate estimate published with the data (three replicates per condition). A site is called significant in a condition when q < 0.05. A q of 0 is printed as 0: it is the value in the source, meaning smaller than the resolution of the estimate.

Colouring the proteins

By default the proteins are tinted by subunit. After a table with per-protein values has been loaded, Colour the proteins by appears and each protein can be coloured by one of its columns instead; the site colouring stays independent, so a protein-level measurement and the site-level response can be read together. Protein values use green scales (purple to green when they span zero) to keep them apart from the site colours. See colour scales for your data.

Your data on the sites

When a loaded table has rows with a position, its columns appear as additional site colourings named Your data: column. Sites without a value in that column are dark grey. See Your data.

The four environments

What the environment tells you

An exposed lysine, and one in an unresolved mobile segment, can be reached on the particle. A lysine whose nitrogen is enclosed by rRNA or by neighbouring proteins cannot carry SUMO in the particle as it is modelled here; a modification found on it points to the protein outside this particle — free, during assembly or disassembly — or to another state of the ribosome. The other structures show some of these states: the line Elsewhere of the detail card gives the environment of the same lysine in each of them.

The detail card gives the accessible surface of NZ twice: in the structure shown and in the protein taken on its own. A lysine that is exposed in the free protein and enclosed in the ribosome is covered by the assembly.

How it is computed

From the atomic model, without water, ions and ligands. Accessible surface: a probe of 1.4 Å radius rolled over all protein and RNA atoms of the structure, partners included (FreeSASA). rRNA contact: the smallest distance from NZ to any rRNA atom. "In the free protein" uses the atoms of that one protein chain only, in the conformation it has in the ribosome.

Limits

Each structure is one state of the ribosome, prepared and selected for a particular question; surfaces that are free in it can be occupied in other states, and flexible regions move. The mature 80S is vacant: no mRNA, tRNA or factors. The thresholds (4 Å, 10 Ų) are conventions; the detail card shows the underlying numbers. SUMO is a protein of about 11 kDa, far larger than the probe: "exposed" means reachable by solvent, which is necessary for modification but does not prove that the conjugation machinery fits.

What it measures

An exposed lysine can be reached by water. SUMO is a protein of 93 residues, about 11 kDa, and the lysine environment does not say whether it fits. Room for SUMO asks exactly that. Picture a SUMO held by its tail at the lysine and turned in every direction, closer and farther: in how many of these positions does it bump into nothing? The folded core of SUMO2 is placed at the lysine in such ways, and the share of placements that overlap nothing is the measure.

For a sense of scale: a lysine on the open surface of a protein reaches about 30 to 40 % (RanGAP1's SUMO site alone reaches ); half of all placements would need the lysine on a stalk. On a ribosome, with neighbours on most sides, a few per cent is already a comfortable fit. An exposed lysine can have no room when its nitrogen is open to water at the bottom of a cleft, and a lysine classed buried can have room when its nitrogen sits at the surface next to what it touches.

The share is computed for every structure of the menu, so one site can have room in an assembly intermediate and none in the mature ribosome. The detail card of a site gives the share, and the colouring Room for SUMO paints the classes.

How it is computed

The core of SUMO2 (residues 17–88, PDB ) is rigid. The five residues after it and the bond to the lysine form a flexible tether. A placement fits when three things hold: the last residue of the core lies to Å from the lysine's nitrogen, the span of the five tail residues from compact to stretched; the tether has a way out, that is, there is room for the glycine bonded to the nitrogen and the straight line from there to the core is free; and no backbone or Cβ atom of the core comes within Å of an atom of the structure. The placements are tether directions × tether lengths × orientations of the core, the same set at every lysine, so that the shares compare between lysines and between structures.

The test is lenient: side chains of SUMO are not tested, because they can rearrange. A class none is therefore the firm statement; a class room says that the backbone of SUMO has space, which is necessary, not sufficient.

The set of placements is finite. The class boundaries are chosen so that this hardly matters: with four times as many placements the class room stays the same for all but a handful of lysines, and a few per cent of the lysines classed none move to tight. A single lysine near a boundary can be on either side; the shares themselves are the better guide there.

The model was tested on a crystal structure of SUMO2 conjugated to RanGAP1 (PDB 3UIO): the SUMO seen there passes the test, its tether measures Å, and of the placements fit with the enzymes of that structure bound, on RanGAP1 alone.

What takes the room away

The same placements are tested four times, each time with less of the structure: as it is; without the partners (factors, tRNA, mRNA); without the other ribosome of a disome; and on the lysine's own subunit alone. Each step can only add room, so the four shares in the detail card never decrease from one to the next; where a share does not change, the thing removed was not in the way, and the card leaves it out. When a site has no room in the structure as it is and gains room once something is taken away, that something is named: room taken by its partners, by the other ribosome, by the other subunit. If more than one is in the way, the first of these in that order whose removal gives room is named. A site that has no room even on its own subunit is enclosed by its own subunit, and nothing is named. The third menu above the table filters by it.

A site whose room the other subunit takes can carry SUMO on a free subunit but not in the assembled ribosome: a SUMO there stands in the way of the subunits joining, or has to be removed first. The same holds for the factor or the neighbouring ribosome.

SUMO at the sites

The control SUMO at the sites under the colouring draws the SUMO2 core at many sites at once, each at a placement that fits (or, where the room is taken, where it would sit without what takes it); a site without any placement gets none, so what is drawn is what the structure can physically carry. Selected site is the default: a SUMO only where the detail card asked for one. All with room draws one at every site the table shows that has a placement, so the filters above the table restrict it. Strong sites keeps the sites above a threshold of a colour scale — the slider, 70 % by default — as the sites most likely to carry the modification. The scale is the colouring shown when it is a fold change or a Z-score (for MG132 and heat shock, 70 % is a log2 fold change of 5.6 among the significant sites; for a Z-score, a value above 0.6); with a class colouring, which has no scale, the MG132 fold change is used and the slider says so. At most SUMOs are drawn, the strongest first. The legend entry counts them and, as every legend entry, hides and shows them. The link carries the mode (sumos).

Showing a SUMO

In the detail card of a site with a possible placement, Show a SUMO here draws the core of SUMO2 in green at one of the placements that fit, one from the middle of the possible directions. It is an example, not a prediction: with sites and thousands of possible placements each, the picture shows where a SUMO can be, not where it is. The tether is not drawn, so the SUMO sits up to Å from the lysine.

For a site whose room is taken, the button draws the SUMO where it would sit without what is in its way; it then overlaps that partner, ribosome or subunit in the picture. One SUMO is drawn at a time; choosing another structure removes it.

The free protein

A lysine with no room on any particle may have room on the protein alone, before it is built into a ribosome or after it is released. For every ribosomal protein the predicted model of the protein by itself () stands in for the free protein, and the same placements are tried on it. The detail card gives the result — room and the share of placements that fit on the free protein — together with the confidence of the model at that lysine (pLDDT: very high ≥ 90, confident 70–90, low 50–70, very low below 50). A low confidence usually means a flexible segment: room there is real but the exact shape is not.

Show the free protein in the detail card replaces the ribosome with the model of that protein, with the same colouring of its sites; the banner at the top offers to colour the protein by confidence instead of its subunit tint, and Back to the ribosome returns. On the model, too, a SUMO can be drawn at the selected site, at a placement that fits there; the tether and the core are the same as on the ribosome. The standard views do not apply to a model, which has its own frame of reference. The colouring Room on the free protein paints every site of the ribosome with its class on the free protein, so that the two can be compared on the particle itself.

The models are predictions of the protein alone, made for the whole sequence, with the tails that no structure resolves. They say where a SUMO fits on the isolated protein, not whether the protein is modified there; and a protein that is never free in the cell, because it is folded into the ribosome as it is made, is a case the model does not know about.

Room for SUMO computed in your browser

For a protein from the AlphaFold DB (Any protein from the AlphaFold DB) or an entry of the PDB (Any entry of the PDB) no analysis has run beforehand, so the viewer computes the room for SUMO itself, in the browser, as soon as the structure is on the screen. It is the same calculation as for the ribosome structures (How it is computed) with the same set of placements — the core of SUMO2 from PDB , tether directions × lengths × orientations, placements at every lysine — taken over from the analysis, so that the shares compare with those on the ribosome and on the free ribosomal proteins; on the ribosomal proteins the two agree to the last digit.

The calculation runs in worker threads, so the picture stays responsive. A bar at the top left counts the lysines done; until it is finished the colouring Room for SUMO and the control SUMO at the sites are disabled, and the table and the detail card say computing. A protein of a few hundred residues takes under a second, a large one a few seconds. The result is the share of placements that fit and the class (room, tight, none) for every lysine of the model — in the table, in the detail card of a site and in the tooltip of any lysine in the structure — and, where placements fit, a placement to draw a SUMO at. In a PDB entry the chains are kept apart, as the pipeline keeps the ribosome's partners and subunits apart: each lysine gets the share in the entry as it is and the share on its own chain alone.

The model is the current version in the AlphaFold DB, named in the banner. The coordinates come from the AlphaFold DB to your browser and the calculation never leaves it; nothing is uploaded.

Limits

The structures are rigid here; parts that move in the cell can make room or take it. Segments without coordinates are not there to collide with, so room near a mobile tail is overestimated. One SUMO is placed; a chain of several needs more space, so none for one SUMO holds for a chain as well. Room for SUMO does not mean room for the enzymes that attach it.

The idea

A SUMO can sit at a lysine only while there is room for it. The structures of this viewer are stages of the ribosome's life — the protein on its own, the assembly intermediates, the mature subunit alone, the mature ribosome, the translating ribosome — so for every site the room for a SUMO can be read off at each stage, and the pattern over the stages, the site's window, says when the modification is sterically possible. A site whose window closes with maturation can carry SUMO during assembly but not on the finished ribosome; a site with room only on the free protein can carry it before assembly or after release; a site with room at every stage says nothing about the stage. The window is a possibility, not an observation: it says when a SUMO can be there, not when it is.

The colouring When a SUMO fits paints the sites by their class; the detail card shows the class, one chip per stage with the room there (hover a chip for what takes or encloses the lysine), and the stage at which the protein is first seen in a structure. The filter menu offers the classes and the window of assembly.

The stages

At a stage with several structures the best share of placements counts. The room is read as room, tight or none; taken means room only once the partners of that structure (assembly factors, tRNA, mRNA, the collided ribosome) are taken away; not modelled that the lysine has no coordinates there; absent that the protein is not part of the structure — a protein that joins late is absent from the early precursor.

The classes

The rules, in order: a lysine not modelled in any particle is a flexible segment; with room on the mature ribosome it is throughout, translation state (if a translating structure covers it) or mature only (if the precursor it is part of does not offer room); without room on the mature ribosome it is free subunit only, precursor only, translation state (room only while translating), free protein only, or none found. For a protein that only the decoding 80S resolves (the L1 stalk), that structure stands in for the mature ribosome.

Reading the window

The classes free protein only, precursor only and free subunit only together are the window of assembly: sites that can carry SUMO only before the ribosome is complete (or after it comes apart). They are the candidates for a role in ribosome biogenesis, and the Statistics window lists them with the stage that closes the window and what takes the room there — an assembly factor named in a chip's tooltip is a site whose SUMO and that factor cannot be there at the same time. The other classes are sites of the finished ribosome: constitutive (throughout), opened by maturation (mature only) or tied to translation (translation state).

Whether a site in the window of assembly is modified during assembly, or on the surplus protein that never assembles, the structures cannot tell; the MG132 response, which raises the unassembled pool, and the heat-shock response are tabulated by class for that reason.

Limits

Steric only, and rigid: the window says where a SUMO fits, not where SUMO is found. Three precursor structures stand for assembly, two of the large subunit and one late one of the small; the early assembly of the small subunit is not covered, so 40S sites without room on the late pre-40S and the mature ribosome fall to free protein only where an early pre-40S might offer room. Room near a segment that is not modelled is overestimated, and a site that is not modelled anywhere is called flexible without further knowledge.

Columns

Filtering and sorting

The search box keeps the sites whose name contains the text (RPS3, RACK1 K2). The first menu restricts to a subunit or one of its regions, the second to one environment, the third to sites in an MG132 or a heat-shock hotspot (any, or one in particular), at the subunit interface, near a partner, on one ribosome of the disome, by room for SUMO or by what takes that room, changed against the mature 80S, or with another modification recorded; it offers what applies to the structure shown. The significance switch keeps the sites that change significantly (q < 0.05) with MG132, with heat shock, with either, or with both. The counter shows how many sites remain.

The Show buttons of the Statistics window set these filters and the matching colouring for you.

Click a column header to sort by it; click again to reverse. Number columns start with the largest value.

Selecting a site

Click a row. The structure focuses on the lysine — a close-up in which what lies in front of it fades away (Overview at the bottom right of the structure brings the whole back; see Focus and overview) — and outlines it, and the details appear below the table. Rows in grey are sites that are not modelled in the structure shown; they have details but no position to focus on. The line Elsewhere of the details tells whether another structure resolves them.

The site details

What is compared, and where the numbers come from

A SUMO site is a lysine. Every question in the Statistics window is therefore asked against the other lysines of the same proteins: sites on proteins against lysines of those proteins that are not sites.

The numbers are computed once, by the analysis that produces this viewer's data, from the published site catalogue, the structures and UniProt. The window shows these stored results; it does not recompute them when the table is filtered or a table of your own is loaded.

The window follows the structure shown. For the mature 80S it holds the five analyses described in the entries below; for the other structures it holds what that state adds: see the other structures.

Statistics of the other structures

This structure counts, per ribosome, the ribosomal proteins present, the sites listed and those with coordinates, their environments, and the sites covered by a partner; it names the sites that have coordinates here and not in the mature 80S. For single particles it gives the fit of the superposition on the mature 80S (root-mean-square distance of the rRNA phosphorus atoms used).

A click on a site in any table of the window closes it and selects that site.

Near the partners

For each kind of partner in the structure, the table counts the SUMO sites and the other lysines of the same proteins whose NZ lies within Å of it, among those with coordinates, and gives the odds ratio with its 95 % interval and p-value from Fisher's exact test. q adjusts the p-values of all such tests across all structures (Benjamini–Hochberg). In the disome each ribosome is tested separately. The second table lists the sites concerned.

A lysine next to a factor in a structure can be covered by that factor, or can be where the factor is modified from; the distance alone does not tell which. The counts are small, and the sites of one protein (several of them on RACK1 in the disome) are not independent.

Against the mature 80S

For the assembly intermediates. The lysines whose NZ is enclosed in the mature ribosome are followed into the precursor: exposed there, still enclosed, not modelled, or on a protein that has not yet joined the particle. The first table gives these counts for the SUMO sites and for the other lysines of the proteins of that subunit; the line below it tests, among the lysines resolved in both, whether the SUMO sites are exposed more often than the others (Fisher's exact test). The last table lists the SUMO sites that the precursor exposes.

An enclosed site that is exposed in a precursor could be modified before the ribosome is complete. A precursor structure is one snapshot of a pathway with many intermediates, and the earliest ones are not covered.

Modified lysines against the other lysines

For a yes/no property (modelled, in contact with rRNA, NZ enclosed, at the subunit interface) the table gives the share of SUMO sites and of other lysines that have it, and the odds ratio with its 95 % interval from Fisher's exact test. For a measured quantity (accessible surface of NZ) it gives the two medians and the shift between the groups (Hodges–Lehmann estimate, Wilcoxon rank-sum test).

q is the p-value adjusted for the six questions of a comparison (Benjamini–Hochberg). Rows with q < 0.05 are tinted. The second table asks the same questions of the MG132-responsive sites against the sites that do not respond.

Subunits and regions

The regions are derived from the structure. The 40S head is the stretch of 18S rRNA that has the fewest contacts with the rest of the 18S — nucleotides –, the 3′ major domain — and a 40S lysine belongs to the head when the nearest 18S nucleotide lies in that stretch; the other 40S lysines form the 40S body. A 60S lysine within Å of the 5S rRNA belongs to the central protuberance, the others to the 60S body. A lysine is at the subunit interface when its NZ is within Å of the other subunit.

The table gives, per region, the share of its lysines that are SUMO sites (with a 95 % Wilson interval), the odds ratio against all other lysines, the number of responsive sites and their median fold change. The subunit table compares 40S and 60S directly. Because the sites of one protein are not independent, the fold change is also tested with one value per protein (its median).

Spatial clustering

The test counts the pairs of sites whose NZ atoms are within Å of each other and compares the count with random draws of as many lysines. Three ways of drawing answer three questions:

"Expected" is the mean of the draws, with the range that holds 95 % of them. p is the share of draws that reach the observed value. The mean distance to the nearest other site and the number of sites in hotspots are tested the same way.

Hotspots

A hotspot is a group of at least MG132-responsive sites in which every site has another within Å (NZ to NZ). Hotspots are numbered by size. Show colours the sites by hotspot, lists the sites of that hotspot in the table and turns the structure to it.

Hotspots describe where the modification is densest. Whether there are more sites in hotspots than chance places there is the last line of the clustering table.

Other modifications

From the cross-link and modified-residue records of UniProt for the proteins of the structure. The table gives, per modification, the number of lysines that carry it, how many of them are SUMO sites, and the odds ratio against the lysines without it. UniProt's own SUMO records are not used here: they largely come from the same kind of experiment as the site catalogue.

A lysine found modified in one proteomics experiment is more likely to be found in another, whatever the modifier. An overlap therefore shows that the same lysines are used; it does not show that one modification depends on the other.

Room for SUMO

The first table counts the classes room, tight and none among the MG132-responsive sites, the other sites and the other lysines of the same proteins, for lysines with a resolved side chain in the structure shown. The second gives the classes of the sites by the environment of their lysine: an exposed lysine need not have room for a SUMO. For the mature 80S a third table tests the differences (Fisher's exact test). The last table lists the sites that have room only when the partners, the other ribosome or the other subunit are taken away, with the share of placements that fit with and without it; a click on a row selects the site.

Definitions and the test of the model: how it is computed.

When a SUMO fits

For the mature 80S the window adds the classes of When a SUMO fits: how the sites distribute over the classes against the other lysines of the same proteins (Fisher's test per class, q by Benjamini–Hochberg; a χ² test with simulated p over all classes), the MG132 and heat-shock response and ubiquitin at the same lysine by class among the sites, the window of assembly against the other sites, and the sites of the window of assembly with the room at every stage.

On the free protein

For the mature 80S the window adds the same questions on the AlphaFold DB models of the proteins alone: whether SUMO sites have room on the free protein more often than the other lysines, whether the lysines that have no room on any particle do, and whether the models are less confident (pLDDT below 70, that is, flexible) at SUMO sites than at other lysines. The next table counts, for the lysines without room on any particle, their room on the free protein, and the last lists the SUMO sites among them, each with a Show button that opens the model of the protein at the site. See The free protein.

Limits

Sites on one protein are not independent observations, so a single heavily modified protein can carry a result; the tests "with proteins as units" guard against this where it matters. Lysines that mass spectrometry cannot detect are counted as unmodified. Each structure is one state of the ribosome. A p-value near 0.05 in a table of many tests is weak evidence.

Where your file goes

Nowhere. Load a table… hands the file to this page, which reads it inside your browser. It is not sent to a server, not written to the browser's storage, and it is gone when the page is reloaded or closed. The page itself only downloads its own files and the structure from the PDB.

File format

A text table with one header row: CSV, TSV or semicolon-separated; the separator is detected. Decimal points or decimal commas. Drag the file onto the Your data box or use the button. Download an example.

How rows are matched

A row without a position is a protein value and colours the whole protein. A row with a position is a site value and colours that lysine if it is one of the listed SUMO sites. After loading, the box reports how many proteins and sites were matched, how many rows name a protein that is not in this structure, and how many positions are not among the listed sites. When another structure is chosen the table is matched again; factors of the structure can be coloured as well, by gene symbol or accession. In the disome a value colours both copies.

One table may contain both kinds of rows. If a protein or site occurs twice, the last row counts.

Colour scales

A column counts as numeric when at least 80 % of its filled cells are numbers. Proteins use green scales, sites the sand-to-indigo and indigo-to-rose scales of the built-in colourings. Proteins without a value keep their subunit tint; sites without a value are dark grey. The legend under each control shows the scale in use.

Where your values appear

In the structure and its legend; as an extra, sortable column of the site table (the site's own value, or its protein's); and in the detail card of a selected site, which lists all columns. Remove takes the table out again.

Limits

One table at a time. Site values are shown for the listed SUMO sites only, not for other lysines. Proteins that are absent from the structure shown cannot be coloured; which proteins a structure lacks is under what is not shown.

Sites and quantification

Hendriks IA, Lyon D, Su D, Skotte NH, Daniel JA, Jensen LJ, Nielsen ML. Site-specific characterization of endogenous SUMOylation across species and organs. Nat Commun 9:2456 (2018), Supplementary Data 1, CC BY 4.0. Endogenous SUMO2/3 sites of HEK293 cells in control, after MG132 and after heat shock, in triplicate; Z-scores, log2 fold changes and q-values are used as published. The viewer shows the sites of the cytosolic ribosomal proteins that are part of the structure shown.

Structures

All structures are cryo-EM models of the human ribosome from the PDB; the table under Choosing a structure lists them with their PDB codes, which link to the entries and their publications. The reference is PDB 8QOI: human 80S ribosome at 1.9 Å (Holvec S et al., Nat Struct Mol Biol 31:1251, 2024); 75 proteins and four rRNAs. Your browser fetches the coordinates from the PDB each time; they are not stored with this page.

Which chains are ribosomal proteins, which are factors, and which RNA is rRNA, tRNA or mRNA is read from the PDB entry. Chains for which an entry records no UniProt accession are identified by their sequence against the proteins of the reference.

Placing sites on the structure

Positions are UniProt sequence positions. They are converted to residues of a structure through the alignment that the PDB entry records for each chain, and every placed site was checked against the published sequence window (the seven residues on either side of the lysine agree with the structure).

What is not shown

A structure lists the sites of the proteins it contains. The mature 80S lacks the mobile stalks (uL1/RPL10A, uL10/RPLP0, P1, uL11/RPL12, uL16/RPL10) and the paralogues RPL7L1 and RPL26L1; the decoding 80S resolves RPL10A, RPL12 and RPL10, and the pre-60S particles hold RPL7L1. The P-stalk proteins RPLP0 and RPLP1 have no side-chain coordinates in any of the structures, and RPL26L1 is in none. Not listed anywhere: lysines in the ubiquitin moieties of RPS27A and UBA52, which are cleaved off before these proteins enter the ribosome, and sites of non-ribosomal proteins, the factors included.

Software

Drawn with Mol* (Sehnal et al., Nucleic Acids Res 2021) from MolViewSpec scenes. The files data/<PDB code>/scene_*.mvsj also open in the stand-alone Mol* viewer. Structural annotation with gemmi and FreeSASA. This page is one HTML file without a build step; licence MIT.