Comments Concerning Researchers' Thoughts on Kinesin Nomenclature

(Ordered Chronologically from Oldest to Newest)




SCOTT DAWSON:   Kinesin groups should really be defined by monophyletic groupings based on good bootstrap/or other confidence support. That should alleviate future issues of paraphyletic groups like KinI or KinCs.


BILL SAXTON:   The idea for naming sequence-related groups by Arabic numerals is fine with me. I guess we could discuss which numbers would be assigned to which groups........... Past the existing kinesin I (1) and kinesin II (2), perhaps it would help memories a bit to proceed from N-terminal to C. So N-terminal kinesins would be 1-7, central motor domain kinesins would be the next number, and C-terminal motors would be the last two. For order within those two multifamily classes [i.e., N-terminal and C-terminal], ?????????? alphabetical order from their current names?


CAROLYN LAWRENCE:   Based upon Bill Saxton's comment (above), I should ask whether researchers want to divide the C-Terminal group into two individual families or whether they should stay lumped together as a single family. Could some of you please report your opinions on whether we should have one or two C-Terminal families?


RON VALE:   Although I was not at the ASCB meeting, it seems as though a consensus is building to name with Arabic numerals (e.g. kinesin 1, kinesin 2), which is fine with me. It would make sense to call conventional kinesins kinesin 1 and the heteromeric, KIF3 type kinesins kinesin 2, largely because people are using these terms already. The only other possible assignment that could make sense is kinesin 4 for the BimC class, since they appear to be tetrameric. Perhaps KIF1/Unc104 could be kinesin 3, since this would keep the "cargo transporters" together.

Bill's proposal of making the C-terminal and middle (internal) domain kinesins follow the N-terminal ones in number also seems reasonable. The only potential problem is if people feel that a future N-terminal class should join, then it would get stuck at the very end, thereby splitting the N-terminal motors. However, this is not terrible. Regarding Carolyn's question on the www, I think that we could make subclasses of the C-terminal motors by calling them A and B (e.g. kinesin 9A). This could be done for other classes later, if people felt the need to split them. However (as stated below), simpler is better at the beginning, as it will be hard enough for people to remember 10 things.

I would add another term of "orphan kinesin" to encompass all kinesins from individual organisms that are not easily classified by sequence into the 10 accepted classes. The term "orphan" is accepted in other field (e.g. "orphan nuclear receptors"). If there is a compelling argument for making a new class, it can be added later (e.g. kinesin 11). However, limiting the classes of kinesins to 10 for the time being seems to make sense. These classes fall nicely into phylogenetic families and account for the bulk of the literature. Most importantly, it will be easier for the field to digest and remember (the nomenclature is meant to help the broad biology community, not kinesin researchers who can survive with complexity). In my opinion, the myosin nomenclature, while good, has suffered by the large number of classes, many of which are sparsely populated.

I (probably others as well) have trouble with remembering numbers. So it might be good to augment the numerical classes with an optional, but consensus "common" name. While the "common" name may not be perfect in terms of description of biochemical or cellular functions, it may help to orient readers as to the nature of the motor class, rather than have people commit to pure memorization. If there are "optional" (but agreed upon) "common" names, this might help to simplify the literature in the future. It might be worthwhile for Carolyn to put forth a suggestion for "common" names corresponding to each class.

For example, in a review or paper, one could use the following sentences:

Human Eg5, a member of the kinesin 4 (bipolar) class, is involved in establishing the bipolar spindle.

or simply

Human Eg5, a member of the kinesin 4 class, is involved in establishing the bipolar spindle.


CLAIRE WALCZAK:   I've been reading through the proposals about the kinesin nomenclature and agree that it is a mess, although many creative solutions have been proposed.

I am in favor of any scheme that links together simple numerical categories along with some evidence of function. I agree with what several others have proposed that the numbers be kept to a minimum number of categories although imposing an arbitrary limit of say 10 may not be appropriate especially when one considers the analysis of Dawson on the protist kinesin families.

Although I see Saxton's point about going in order- N to Internal to C-terminal, I agree that if a new family of N-terminal kinesins comes up, this creates a problem because we would have a split in the numerical ordering.

I currently like the current Kin N, I and C, although I do fully understand that I was an awful choice as it is often called Kin One- which gets confused with conventional kinesin. The abbreviation should absolutely by Kin or just use the full kinesin as proposed by Scholey- this is the enzyme family named as it was discovered. Simple arabic numbers would be very good as I agree that roman numerals have their problems in the system.

I agree with Scholey, Ron, Saxton- that some evidence to a family must be given to put it in context. At the present time, I cannot decide whether the Drosophila or vertebrate nomenclature is more confusing- there are just too many numbers to keep anything straight. There are problems with the descriptors being functional or structural and these have been pointed out by others- but there are definitely distinguishing features of the families that should make it possible to come up with some reference that we could all remember.


MANFRED SCHLIWA:   I have nothing to add, except to express my hope that all will agree on one of the schemes. Now is the time. If we (the community) fail to do it now, we will have to live with the mess forever.


TIM MITCHISON:   I personally do not find the myosin nomenclature to be so great, though I agree its better than the current kinesin nomenclature. Using numbers provides no clue as to function or structure, and is difficult for people outside the field to use or remember. It also confuses probably orthogolous relationships, like that between myosin VII and X.

I like the N/I/C nomenclature for kinesins because at least it gives a bit of information about structure.

Maybe something like Vale's suggestion where one could use a traditional name and append a systematic one would be a good compromise?


NOBUTAKA HIROKAWA:   Most of the above debate seems to be centered around which number to assign to which family in the case there is no other way to reflect function or structure in a family name. Will we reach a conclusion and can we resolve the confusion? Even Ron says "It would make sense to call ... KIF3 type kinesins kinesin 2, ... KIF1 could be kinesin 3". So 3 is 2 and 1 could be 3. We have already tried in the past to introduce, with Ron, a numerical family name system, as Ron, John (at the ASCB meeting) and Claire are proposing now. But it did not work. Linking a kinesin with a seemingly irrelevant family number is complicated. As Tim has pointed out, it is difficult to remember (even I fail to remember the numbers of family names I proposed). In addition, a new naming convention with family numbers different from the previous nomenclature will be even more confusing.

If we use an alphabet in place of a family number we could avoid confusing number conversions and the issue of which order to number the families. Also, a one-letter alphabetical code, since never previously used for kinesin families, will be less confusing than re-numbering families. Alphabets can act in place of symbolic numbers and simultaneously represent a "common name". For example, for the kinesin heavy chains, we could name them the "K-type kinesin" for KHC, or "B-type kinesin" for the BimC or Bipolar family, or "U-type kinesin" for the Unc104/KIF1 family. This system compromises for a traditional naming scheme and a systematic one (though the use of "KIF"s would be so much more systematic and somewhat traditional). This scheme was inspired by one of Manfred's ideas. Please refer to my proposal on the "Nomenclature Information Resource" page for details. This system also fits with the idea that the new system should not be biased by anyone's previous proposal.

We can have up to 26 families this way (or 25 if we avoid to use "I" to minimize confusion) . Recent studies using large numbers of kinesin sequences from Scott Dawson and Zac Cande's group, Sharlyn Endow's group and ours agree that there are around 15 families. Ten may be too arbitrary a number and lacks scientific rationale. We are dealing with science, not politics. Nature, not we, should determine the number of families. This should really be based on phylogenetic analysis (as Scott has stated above).

The term "kinesin" is fine by me, since I sense it is more acceptable to you than "KIF". I agree with John that Ron has given these molecules a respectable name.

However, I cannot accept family names containing terms with functional implications even as "common names". Bad examples such as "Chromokinesin" or "Rab6kinesin" are apparent as there are reports with views against the function that these names implicate. Also note that the so-called "Internal kinesin family" contains kinesins that do not have internal catalytic cores and many other families contain kinesins that have a catalytic core close to the middle.

By using alphabets, we do not have to argue whether to start which family name with "N" or "I" or "M" or "C". No need to worry about the order of the family nor about the future addition of the new families, either. It will be easy, if found necessary, to subdivide a family by appending an additional letter or number such as "C1" or "Cb". Since most comments favor either a numerical systematic nomenclature or a "common name" (or founding member) system, I hope the above scheme could set a foundation for a compromise between the two.


CAROLYN LAWRENCE:   Dr. Hirokawa's proposed schema (see the pdf available from the table on the main page) brings one question to my mind: what are the criteria that should be used to deem a group of sequences a "family"? While many of the families proposed in this schema make sense, I question whether the small group named "F" in that schema should be called a family. Shouldn't a family have members occurring in more than one kingdom (in this instance, Animalia)? Surely in more than one class (in this instance, Mammalia)... Could some of you please report your opinions on this sticky wicket?


NOBUTAKA HIROKAWA:   I would like to answer Carolyn's comment and address some issues for all participants of this web session.

As seems to be the consensus, there should be orphans. However, for the "F-type kinesin" family, members derive from protozoans, animals and also chlamydomonas (CrKLP1). Additionally, distinct features in the neck domain support establishment as a family. These kinesins clearly do not belong to any other family.

So, instead of lumping them into the orphan family for now and later announcing that they should form an independent family, I think we should try to avoid future confusion and not put off family definition. A solid definition is what we are trying to achieve.

Future members may be found in higher plants since, of the 600 plus kinesin sequences in public databases, only around one fifth are from plants. More than half are from animals. At this point, we may not have adequate data to support family designation by the presence of kinesins from all kingdoms.

I would also like to point out that the "T-type kinesin family" (KIF3A/B, KRP85/95, KIF17, osm3,...) also lacks members from higher plants. Do you mean to lump them together with other orphans, Carolyn?

We should define families now by objective and scientific methods such as phylogeny. By comparing catalytic core and/or full length sequences, most groups find around 15 kinesin families. This is a purely scientific issue that researchers should discuss and reach an agreement based on valid data.

How to name these families is a highly political issue that researchers in and around this field should participate in. For the new nomenclature to be used widely, opinions from as many researchers as possible should be considered. A biased agreement will not only inhibit wide approval, but also I fear it will not be adopted by researchers even within this field.


KELLY DAWE:   I have been following this debate for years now and have considered all the proposals at length. I'm in favor of a strictly numerical scheme (roman or arabic) followed by a common name. I think new families should designated only when it can be shown that the family exists in two kingdoms, e.g. protozoa and animals, or plant and animals etc. This solves what I see are the two main issues:

1) Carolyn first saw the need for a unified naming scheme when she realized that KinI was being misused. Many if not all of you recognize this as a problem and it should be solved. For this reason, we really cannot safely use a letter designation because KinI will forever haunt us.

2) It seems clear that we need some kind of common name designation, at least in the beginning, since we are all absent minded. Carolyn has put forward a rationale for choosing what those names should be that makes sense to me and many of you agree. I vote we give Carolyn the freedom to choose common names, trusting that she understands the field, the concerns, and the people involved.


*COMMENTS FROM HERE ON WERE RECEIVED ACCOMPANYING BALLOTS FROM THE VOTE.*  


MARK ROSE:   This seems to me to be a very sensible proposal. Each scheme has it's benefits, but I think that there are really only two that are worth serious consideration, the Arabic Number and the Mnemonic Roman. Although there is much to be said for simply assigning arbitrary numbers (loosely respecting some of the prior designations), I think that there ii still some use to keeping some information about the family tied to the family name. For those who have some knowledge of the families, the Mnemonic Roman scheme will at least serve as a cue, whereas the arabic number serves only as an index and conveys no other meaning. On the whole then, I think that this would be more useful to students and more casual perusers of the literature. It seems to me that these are the ones for whom the family designations wil be most useful, the experts will already know from the gene name. However, I do not feel strongly about this and I greatly appreciate the effort that has gone into the proposal. Thus, I have rank ordered my vote.