Restriction enzymes

The Enzymes tab scans the sequence for the recognition sites of a bundled table of about 130 common, commercially available enzymes. The scan runs in the background whenever the sequence changes, on both strands, and handles ambiguity codes in recognition sequences (GRCGYC), enzymes that cut outside their site (Type IIS such as BsaI and BsmBI) and sites that wrap the origin of a circular sequence. You can swap the bundled table for the whole of REBASE — see Importing the full REBASE table below, or add a single enzyme of your own — see Adding your own enzyme.

Reading the list

Each row is one enzyme that cuts the sequence — or, with isoschizomers sharing a row, every enzyme that cuts it identically (see Isoschizomers):

• a tick box that chooses the enzyme: its cut sites are drawn in the views, and the Cloning tab digests with it;
• the name, with +N after it when N other enzymes share the row (hover it for their names);
• the recognition site, 5′→3′ on the top strand (hover to see whether it leaves a blunt end, a 5′ overhang or a 3′ overhang);
• the cut positions: the base after which the top strand is cut (a cut exactly at the origin of a circle is after its last base, so it reads as the sequence's length). Click one to select that recognition site in the views;
• the bands this enzyme alone would give, largest first, as a gel would show them. Fragments within about 15 % of each other are written as one band (2,181 ×2), because that is what you would see. A ⚠ means the lane would hide something — pieces running as one band, or bands under 100 bp that may run off the end — and the tooltip says which. An enzyme that simply linearises the plasmid gets no mark: there is nothing to tell apart, which is not a fault.

The list scrolls inside the panel, so what is below it — the counts, the import link, the fragment sizes — stays where you left it however many enzymes cut. Only the rows on screen are drawn, which is what keeps an imported table of 1,500 enzymes responsive; a browser page search (Ctrl+F) will not find an enzyme that is scrolled out of sight, so use the filter box for that.

Below the list: how many enzymes cut, and how many do not cut at all.

Choosing which to show

• When a document opens, the enzymes that cut exactly once are ticked, so unique sites are on the map at once. If there are more than 50 of them — which happens with an imported REBASE table, where a small fragment can have ninety unique cutters — nothing is ticked instead, because ninety labels is a wall rather than a starting point. The tab then offers Tick the N enzymes that cut once if you want them after all; with a Cuts filter set, it offers those instead — Tick the N enzymes that cut twice.
• Tick or untick individual enzymes. Show listed ticks every enzyme currently in the (filtered) list, Hide all clears them. Show listed asks for a narrower list past 200 enzymes: more labels than that is more than the views can draw.
• Filter by name (Eco) or by site (GGATCC).
• Cuts lists only the enzymes that cut a chosen number of times: once (a unique site to clone into), twice, once or twice, or three times or fewer. A diagnostic digest is the usual reason for wanting more than one — an enzyme that cuts twice gives two bands to check a construct against, where a unique cutter only linearises it. Any number is the default.
• Order sorts the list. _Name_ is the catalogue. _Band separation_ puts the enzymes whose bands are easiest to tell apart on a gel first, and the ones whose bands would run together last, which is how a diagnostic digest is actually chosen. Bands twice as long as each other are as distinct as bands ever get, so past that the order prefers a lane whose smallest band is bright — 500 bp or more on a 1 % gel — over one with a faint sliver at the foot. The ↓ button beside it reverses either order: Z to A, or the lanes hardest to read first (enzymes that do not cut stay at the end).
• Sold by narrows the list to one supplier's catalogue. It appears only after you import a REBASE table, which is where the supplier information comes from.

• Isoschizomers: share a row is described below.

Cuts, Order (and its direction), Sold by and Isoschizomers are remembered between sessions, like the toolbar's view settings: they say what you are generally looking for rather than anything about the document. The name filter is not — it is a question about the list in front of you, not a standing choice.

Cut sites of the ticked enzymes are drawn in the sequence view (a mark with the enzyme name above the bases) and on the map (labels around the ring), and they are included in File ▸ Export figure…, where the map's Restriction enzymes toggle can leave them off the figure.

Choosing a diagnostic digest

Set Cuts to _Twice_ and Order to _Band separation_. The enzyme at the top gives the two bands easiest to read, which is the one to run. On pBR322 that is HincII, at 3,254 + 1,107 bp, with BstAPI and DrdI (3,948 + 413) just behind it; BtsI also cuts twice, but its two cuts are 20 bp apart, so it is at the bottom with a ⚠ and one band you could see.

When no single enzyme will do, Double digests under the list offers the five best pairs of the enzymes listed, judged by the same rule on the digest with both — on pBR322, EagI + MscI at 3,854 + 507 bp. It appears while the list is ordered by band separation, pairs only enzymes that cut three times or fewer, and follows the other filters, so _Sold by_ narrows the pairs to what you can buy from one place. A pair whose second enzyme cuts only where the first already does is left out, since it is the same digest as one enzyme alone. Tick both ticks that pair and nothing else, and the gel then shows it beside each enzyme alone. With a big imported table only the 120 enzymes that cut least are paired; narrow the list to pair the others.

A partner for one enzyme

Often the question is narrower: _what do I cut with EcoRI to check this?_ Set Pair above the double digests to _EcoRI with a partner_, and the five best partners for it are listed, looked for among every enzyme that cuts three times or fewer — the name filter and the Cuts filter do not narrow them, only Sold by does. With exactly one enzyme ticked, the section offers Find a partner for it in one click. Set Pair back to _any two listed enzymes_ for the pairs of the list.

Tick a candidate and the Fragments from ticked enzymes section below the list says what the whole digest would look like — several enzymes ticked together give a lane none of their own rows can predict.

A double digest beside the single ones

Tick two enzymes (or three) and the gel draws a lane for each of them alone and a last lane, Both (or All 3), for the digest with all of them — the way a double digest is run at the bench, so a band can be read off against the lanes that explain it. A band in Both that is in neither single lane is the piece between the two enzymes' sites. The sizes are written beside the combined lane; hover a band in any lane for its size, and the line under the gel lists what each enzyme gives alone. Tick four or more and the gel goes back to one lane for them all.

Isoschizomers

Enzymes from different organisms often recognise the same site and cut it in the same place — BamHI, BstI and a dozen others all cut G^GATCC. They give the same fragments with the same ends, so at the bench they are one choice whatever the label on the tube. With Isoschizomers: share a row ticked (the default), they are listed as one row: an imported REBASE table of about 1,550 enzymes becomes about 400 rows.

• The row carries the best-known name: an enzyme from the bundled table if there is one, otherwise the one sold by most companies. A name you searched for, or one you ticked, takes its place, and with Sold by set only that supplier's enzymes can name the row.
• Ticking the row ticks the enzyme named on it, and unticking it unticks every enzyme in the row.
• When a document opens and its single cutters are ticked, one enzyme per row is ticked, so one cut is drawn with one label rather than twelve.
• The count under the list still counts enzymes, and says how many rows they take.

Enzymes that recognise the same site but cut it differently — _neoschizomers_ such as SmaI (CCC^GGG, blunt) and XmaI (C^CCGGG, 5′ overhang) — keep separate rows, because the ends they leave are what a cloning plan depends on. Untick share a row to list every enzyme on its own.

Hiding the cut sites without losing the choice

Cut sites in the toolbar, next to Complement and Translations, hides every drawn cut site at once — useful when a map with a dozen labels is too busy to read. The ticks stay exactly as they are, so switching it back on restores the same set; no need to untick and re-tick a carefully chosen list. The toggle is remembered between sessions (see Viewing and selecting); the ticks are not, they start again at the single cutters of each document you open (or at nothing, with a big table).

While it is off:

• the sequence view, the map and File ▸ Export figure… draw no cut sites;
• the Fragments from ticked enzymes list below and the Cloning tab's digest are unchanged — they follow the ticks, not the toggle.

The Enzymes tab says so while the sites are hidden, and offers a Show cut sites link to bring them back.

Fragments

Fragments from ticked enzymes draws the lane a complete digest with all ticked enzymes would give (with a lane for each enzyme alone beside it when two or three are ticked), beside a ladder — 1 kb, or 100 bp when everything is small, unless you choose one (see The gel below). A short band is drawn faint, because a stain binds DNA by mass and a 200 bp band really is faint next to a 4 kb one. Click a band to select that piece — in a single enzyme's lane, the piece that enzyme alone cuts — in the sequence view and on the map; where two pieces run together, the click takes the larger one, and the band says so when you hover it.

Under the lane are the fragment sizes themselves, largest first — on a circular molecule, n cuts give n fragments; on a linear one, n + 1 — and then On a gel, which says in words what the picture cannot: which pieces would run together, and whether any are too small to stay on the gel.

The gel

Under every drawn gel — here and in the Cloning tab's PCR — two menus say what it is a picture of:

• Agarose: 0.7, 1, 1.5 or 2 %. A thinner gel spreads out large fragments and loses small ones off the end; a thicker one the reverse. The choice changes the drawing and everything judged from it: the ⚠ on a row, the Band separation order and the double digests offered.
• Ladder: chosen to fit what is run (1 kb, or 100 bp when everything is under 1.5 kb), or always the 1 kb, the 1 kb Plus (both scales in one lane) or the 100 bp one.

Both are remembered between sessions.

The gel is a rule of thumb, not a simulation: band positions are calculated from the length, and two fragments within about 15 % of each other are called one band. Each percentage has a range it separates well — 0.7 % from about 800 bp to 12 kb, 1 % from 500 bp to 10 kb, 1.5 % from 200 bp to 4 kb, 2 % from 100 bp to 2 kb. More than one fragment above the top of it will compress near the well, and anything under a fifth of the foot of it (100 bp on a 1 % gel, 20 bp on a 2 % one) may run off the end. Your gel may do better or worse.

For the fragments themselves, with their ends and the features they carry, and to join them into a new construct, use the Cloning tab.

How to check whether an enzyme is a unique cutter

1. Open the Enzymes tab.
2. Type the enzyme name in the filter box.
3. Look at the number of cut positions in its row. Set Cuts to _Once_ to list all unique cutters at once.

Importing the full REBASE table

The bundled table is deliberately small: the enzymes most people clone with. For everything else — and for supplier and isoschizomer information — import REBASE, the restriction enzyme database Richard Roberts and colleagues have maintained since 1975.

PlasmidPop cannot download it for you. REBASE files are copyright "all rights reserved", so we have no right to ship them, and rebase.neb.com does not allow a web page to fetch them. You download the file; the app reads it.

1. Open the Enzymes tab and click Import a REBASE table… at the bottom of the list.
2. Follow Download withrefm from REBASE. The file opens as plain text in a new tab — save it with Ctrl+S, or right-click the link and choose _Save link as_. It is about 4 MB.
3. Back in PlasmidPop, click Choose file… and pick what you saved (usually link_withrefm.txt), or drop the file onto the panel. The picker asks for text files; if your browser saved it without an extension, switch the picker to _All files_.

The table is read in your browser, kept in this browser's local storage, and never uploaded anywhere. It survives reloads, and the Enzymes tab says which release it is scanning with. Go back to the bundled table undoes the import.

Of the roughly 6,100 records in REBASE, about 1,580 become usable enzymes. The panel reports what it left out: enzymes whose cut position nobody has determined, and the ones that cut at a modified base rather than at a sequence (AbaSI, MspJI and kin), whose REBASE site is a base or two and would otherwise appear to cut everywhere.

A few enzymes are written in REBASE with only their top-strand cut (HauII's TGGCCANNNNNNNNNNN^); PlasmidPop knows the other cut for the ones it has checked. If a later REBASE writes one of them differently, or adds a new one, the import leaves it out rather than guess, and lists it under the summary as a warning. Such an enzyme is worth reporting as an issue.

About 27 of them — BcgI, BaeI, CspCI, BsaXI and kin — cut on _both_ sides of their recognition site, cutting it out on a piece of about 30 bp. Each site of theirs is two cuts, so such an enzyme with one site in your plasmid is listed as cutting twice, draws two cut marks, and in a digest cuts out the short piece carrying the site. The tooltip on its recognition site says so. They are not offered for Golden Gate. A table imported before version 1.2 does not have them; import it again to add them.

Scanning with the whole table takes about 15 ms on a 4 kb plasmid instead of 3 ms, and about 130 ms on a 50 kb construct, in the background.

If you publish work that used it, please cite REBASE: Roberts RJ, Vincze T, Posfai J, Macelis D, _REBASE — a database for DNA restriction and modification: enzymes, genes and genomes_, Nucleic Acids Research.

Adding your own enzyme

For an engineered enzyme, a nicking or homing enzyme the table lacks, or a site you want to treat as an "enzyme" in a Golden Gate design, click Add enzyme… at the foot of the Enzymes tab. Give it a name and its recognition site in REBASE notation:

• G^AATTC — the caret is the top-strand cut; the bottom strand is cut symmetrically (EcoRI, a 4-base 5′ overhang).
• G^AATT_C — the same with the bottom-strand cut written out after the underscore, for a cut that is not the mirror image.
• GGTCTC(1/5) — a Type IIS enzyme: the cuts fall 1 base (top) and 5 bases (bottom) past the 3′ end of the site. Negative numbers count back into the site, and (10/12)CGANNNNNNTGC(12/10) cuts on both sides.
• IUPAC codes (R, Y, N, …) work in either case. A caret after N padding (GGTCTCN^NNNN_) needs the underscore, because the caret alone does not say where the bottom strand is cut.

The dialog shows how the site was read: where the top and bottom strands are cut, the overhang that leaves, and whether the site is palindromic. Save enzyme stores it in this browser, and it is at once in the sites list, digests, Golden Gate and the diagnostic ranking, like any other enzyme. It is read by the same code as the REBASE import, so both mean the same by a notation. Your enzymes survive a reload and an imported table (importing or Go back to the bundled table leaves them alone); use Remove in the dialog to delete one. A name already in the table in use is refused, and a site with too few specific bases to be a real cutter is too.

Notes

The bundled table was typed from supplier catalogues and has no supplier, isoschizomer or methylation information; an imported REBASE table adds all three (hover an enzyme's recognition site to see them). What REBASE calls the methylation site is where the enzyme's _own_ methyltransferase modifies the site — it is not a statement about Dam or Dcm methylation from your _E. coli_ strain, which is the next section.

Dam and Dcm methylation

Most laboratory _E. coli_ strains methylate the DNA they copy: Dam puts a methyl group on the A of every GATC, Dcm on the inner C of every CCAGG and CCTGG. Some enzymes will not cut, or cut poorly, where that methyl group lands inside their site — sometimes only when a GATC or CCWGG is made by the site together with the bases beside it, as with XbaI in GATCTAGA.

For the enzymes New England Biolabs lists as blocked or impaired, the Enzymes tab looks at the bases around each site. Which methylase matters, and how much of the site it must reach, is per enzyme, taken from REBASE's tables of overlapping methylation: BstXI, AlwNI and PflMI ignore a GATC inside their N bases, and SfoI is blocked only when CCWGG methylates both ends of its site, not one. FokI is impaired when a CCWGG ends in the first GG of its GGATG, as in CCAGGATG, the one way Dcm reaches its site. A few enzymes NEB lists as sensitive have no such table in REBASE (BclI, DpnII, BssKI, BcgI, FspI and PhoI); for those, any Dam or Dcm methyl group inside the site counts. A cut position whose site a Dam or Dcm methyl group would fall in, in a way that enzyme is sensitive to, is marked with a small m and written in the warning colour; hovering it says which. The row says how many of its sites are affected — _1 of 2 sites may be blocked by Dam/Dcm methylation_ — and the enzyme's tooltip says it is a sensitive one.

On pBR322, for example, the MscI site is inside a CCTGG and the BspEI site inside a GATC, so plasmid grown in an ordinary strain may not cut there.

Where the DNA was grown

Whether those marks matter depends on where the DNA came from, and the document says so: Grown in, among the list's controls, is one of dam+/dcm+ (an ordinary laboratory strain of E. coli, and the default for any plasmid), dam+ only, dcm+ only, or unmethylated — DNA from a dam− dcm− strain, from mammalian cells, or straight out of a PCR. A product the PCR panel makes is unmethylated from the start.

What it changes is the cutting, not the marking. A site stays marked whatever the host is, because "this would be blocked in a dam+ strain" is worth knowing about DNA that is not in one. But a site the host's methylation does block is left out of:

• the fragments and the gel below the list;
• the double-digest ranking, so an enzyme this DNA would silence is not offered as a partner;
• the Cloning tab's digest, and so the shelf, the ligation and the gel there.

On pBR322 that is the difference the guide's own example turns on: EagI + MscI is a clean two-band digest on paper, and plasmid from an ordinary strain will not give it, because the MscI site is Dcm-blocked.

The Cloning tab says when it has left a site out for this reason, and when every ticked enzyme is blocked it says so rather than claiming nothing is ticked.

A piece of a molecule is that molecule's DNA: a selection exported as a file, a digest fragment put on the shelf or opened as a document, keeps the host of the document it came from, so a fragment of a PCR product stays unmethylated. A ligation, Golden Gate, Gibson or Gateway product is what you would transform and grow, so it starts as dam+/dcm+.

Changing it is an edit like any other, so Undo takes it back, and it is saved with the document: GenBank has nowhere to put it, so it travels as a comment line of ours (PlasmidPop-methylation: ...), the way sticky ends do. A SnapGene file brings its own setting with it — the flags its sequence packet carries — and a file that says nothing is read as an ordinary plasmid.

Cut positions are given on the top strand; the bottom-strand cut and the overhang are used by the Cloning tab.