# What does the ZZE tag in an Chemical Table file refer to?

I have a number of Chemical Table files which include a "M ZZE" tag in the properties block.

For example, in Galactobuxin, I have:
M ZZE 1 31 23

In Rubrobrassicin there is:
M ZZE 3 28 23 36 23 48 23

In R-hexanoyl CoA there is:
M ZZE 1 55 23

The typical format for such tags is that it starts with an M, followed by two spaces. Then there is a three-letter code, in this case ZZE. Other codes include ISO (for isotope) or CHG (for charged atoms in the compound), then a number which indicates how many records on that line. This is followed by the actual data. So
M ZZE 3 28 23 36 23 48 23
would probably mean that there are three sets of ZZE data, being "28 23" "36 23" and "48 23". In every record I have the second number is 23.

The ZZE tags seem to occur only for compounds containing ribose rings, and the first number is the index of one of the bonds inside the ribose ring. Specifically, of three three bonds adjacent bonds with a stereochemistry marker, it always refers to the one in the middle:

I'm guessing the 23 is some property of the bond: so in rubrobrassicin the 28th, 36th and 48th bonds (there's one, and only one, ZZE record for each of the ribose rings in rubrobrassicin).

Does anyone know what the ZZE property in a chemical table file refers to, or what it indicates? I can't find any information on the format.

-
Jivlain, i'm sorry but i don't see the ZZE entries in the SDF files on the PubChem files you're linking to. –  Klaus Warzecha Jul 26 '12 at 18:34
Usually, i'd use Z (= zusammen) and E (= entgegen) to denote the stereochemistry at double bonds but that doesn't play a role for the structures you mentioned. Furthermore, to my knowledge, E and Z as described above are not part of the MDL molfile specification. –  Klaus Warzecha Jul 26 '12 at 18:53
@KlausWarzecha: Indeed - I have been sent some SDF files which contain ZZE tags (I just linked Pubchem for convenience's sake). Your zusammen/entgegen suggestion seems quite plausible, however - I had a closer look and noticed that the ZZE tag always seems to refer to the bond in the middle one of the three bonds with a sterochemistry marker. –  Jivlain Jul 26 '12 at 19:29
Could you possibly upload your SDF file somewhere (for example, filedropper.com) and share the link? Also, it may be instructive to know where/how the individual who sent you the SDF file generated them. It may be a designation specific to the program used to generate the file. As @KlausWarzecha indicated, the SDF files on pubchem don't contain the ZZE entries, so we need more information about the origin of the SDF files or the actual files themselves. –  J M Jul 26 '12 at 21:07
Jivlain, i had a look at the CT file specification as provided by Accelrys and i found nothing on Z or E. @JM is right, concerning the origin of the SDF data. Knowing how the files were generated might really help to solve the puzzle. –  Klaus Warzecha Jul 27 '12 at 7:51

The current version of the CTFile spec can be downloaded for free from Accelrys. Based on the files you have, you'll want to look at the "V2000" section.

Short answer: I believe [M ZZE] is a vendor-specific property key, possibly introduced by ACD Labs, given it's the package cited in the header on line 2. The CTFile format spec for V2000 files says nothing about an "M ZZE" property key.

The CTFile spec says that property keys that are not recognized by parsing software are to be ignored. This makes vendor-specific extensions feasible.

You seem to be on the right track in your interpretation. The Galactobuxin entry has this line in the properties block:

M ZZE 1 31 23

Breaking this down:

{M ZZE} is the mystery properties key.

{1} is the number of data entries on the line.

{31} is the 1-based index of either an atom or bond.

Atom 31 is one of the rather boring methoxy groups.

On the other hand, bond 31 is a real problem child:

This file is abusing the meaning of 'wedge bonds' by trying to use them to indicate chair conformation perspective, rather that what they're intended for - flat stereoprojection. My guess is that the file is trying to annotate this bond.

The only question left is - what does {23} refer to?

Bond 23 is unremarkable. Atom 23 is one of the origins of a wonky wedge bond. But it's not clear why it should be associated with bond 31.

My best guess is that the software that created this file knew it was abusing wedge bonds and wanted to include a flag on bond 31 so that software reading the file wouldn't misinterpret.

Given that you're seeing the same {23} in every file, this could be a code to specially process bond 31's stereo flag.

-