Open Close
Davis, T. (2001.4.23). FlyBase error report for CG15284 and CG3474 on Mon Apr 23 02:12:23 2001. 
FlyBase ID
Publication Type
Personal communication to FlyBase
PubMed ID
PubMed Central ID
Text of Personal Communication
From FlyBase-error@XXXX Mon Apr 23  10:12:28  2001
Envelope-to: gm119@XXXX
Delivery-date: Mon, 23 Apr 2001  10:12:28  \+0100
From: FlyBase-error@XXXX
Date: Mon, 23 Apr 2001  02:12:23  \-0700 (PDT)
X-Authentication-Warning: web set sender to FlyBase-error
using \-f
To: flybase-updates@XXXX
Cc: davist2@XXXX
Subject: FlyBase error report for CG15284 and CG3474 on Mon Apr 23  02:12:23  2001
Content-Length: 7675
Error report from Terence Davis (davist2@XXXX)
Gene or accession: CG15284 and CG3474
Release: 1
Missed gene
Comments: I have some annotation information for the genes CG15284 and CG3474.
The predicted gene CG15284 lies on 35A close to (and probably proximal to) the
chromosomal aberration In(2LR)el6 and distal to the deficiency Df(2L)el16
i.e., is most probably between these two breaks. (note that the two
aberrations and CG15284 all map to the same 5.4kb EcoRI fragment)
Deletions which remove DNA proximal to el6 (e.g. the synthetic deletion
el6Left-A379Right) are mutant for the pupal (pu) gene, whereas the el16
deletion (goes proximal) is not mutant for pu. This indicates that the pu gene
is between these two breaks (they are approximately 3 kb apart).
I noticed this putative transcript a couple of years ago and have wondered if
it could be the pupal gene (I realise that Mike Ashburner and I might disagree
about this). The source of disagreement is due to the map position of
Df(2L)el14 (I think I erroneously mapped it: see Davis, T., Trenear, J. and
Ashburner, M. (1990) The molecular analysis of the el-noc complex of
Drosophila melanogaster. Genetics 126: 105-119.). If the published position of
this break is correct then it would delete CG15284 and el14 is not mutant for
pu. As the genomic sequence is available I now think that el14 and el16 are
broken very close to each other and neither would delete CG15284 (of course I
am not in a position to remap this break which would clinch the matter).
If CG15284 is indeed pupal, then CG3474<up> BG:DS06238.4 </up> (this is Ashburner's
guess for pu) is not pupal. This would fit the genetic data as both el14 and
el16 delete CG3474 and neither are mutant for pupal.
Then CG3474 could be a gene which we labelled 'crippled leg' (see Davis, T.,
Ashburner, M., Johnson, G., Gubb, D. and Roote, J. (1997) Genetic and
phenotypic analysis of the genes of the elbow-no-ocelli region of chromosome
2L of Drosophila melanogaster. Hereditas 126: 67-75.). When this 'locus' is
deleted, or affected by aberrations, the surviving flies have a crippled leg
phenotype (e.g. Df(2L)el16/Df(2L)b83d29a which only homozygously deletes
CG3474 is pupal+ elbow+ crle-). There are as yet no point mutations for
'crle'. The existence of this gene relies on the phenotype from overlapping
If CG15284 is pu and CG3474 is 'crle' then the molecular data fits the genetic
data better than assuming that CG3474 is pupal.
Having stated all the above I have no actual proof as it depends upon My
having mis-mapped the aberration el14 (I am fairly sure I did as the EcoRI
fragment I used for the mapping (taken from the genomic sequence) is clearly
at odds with my published position). On the other hand if CG3474 is pu then
homozygously deleting this region does not result in a pupal phenotype? This
is despite the observation that heterozygous combinations of deletions can
indeed be pu- (e.g. Df(2L)A400/Df(2L)b83d29a)
The genetics to support this are:
Df(2L)b83d29a breaks proximal to CG3474 and deletes this locus (is pu- crle-)
Df(2L)b83d29a heterozygous with Df(2L)el14 or Df(2L)el16 is pu+ el+
crle-<up>would delete only CG3474</up>
Df(2L)b83d29a heterozygous with Df(2L)fn2 is pu+ el+ crle- (fn2 is possibly
broken in the promoter region of CG3474: this has a weak crle phenotype)
Df(2L)b83d29a heterozygous with Df(2L)TE35BC8 is wild-type (BC8 is broken just
proximal to CG3474 and fn2)
Df(2L)b83d29a heterozygous with Df(2L)ARR1 is wt (ARR1 breaks proximal to fn2
or BC8)
Df(2L)b83d29a heterozygous with deletion el6Left-A379Right is pu- crle-
<up>deletes both CG15184 and CG3474</up>
Genetic data from: Davis, T., Ashburner, M., Johnson, G., Gubb, D. and Roote,
J. (1997) Genetic and phenotypic analysis of the genes of the elbow-no-ocelli
region of chromosome 2L of Drosophila melanogaster. Hereditas 126: 67-75
Thus the 'crle' gene is between the breaks of Df(2L)el16 and Df(2L)TE35BC-8
and homozygous deletions of this region are semi-lethal with the survivors
having a crippled-leg phenotype.
The locus CG3474 (crle?) encodes a putative pupal cuticle protein
The gene CG15284 (pu) has a low degree of homology to the 3' end of the human
mucin genes and the growth factor for human Norrie's disease (see below). The
domain of homology is a cystine-knot domain. I appreciate that the overall
homologies are low, but all the conserved residues required for the
cystine-knot are 100% conserved (see Meitinger et al., 1993, Molecular
modelling of the Norrie disease protein predicts a cystine-knot growth factor
tertiary structure, Nature Genetics, 5: 376-380). The cystine-knot is required
to fold in a specific way and to cross-link units. Meitinger et al., state
that the overall homologies of this class of proteins is 10-30%, the essential
requirement is for the cystine-knot domain cys residues (the pu homology is
greater than 10%).
The putative pupal protein has reasonable homology to the cystine-knot domain
and all the required cysteines are conserved. In addition the pupal protein
would have a good signal peptide indicating an excreted protein. If this match
is more than coincidence, then pu could be a Drosophila version of the
cystine-knot and thus be a homolog of a member of this growth factor family.
I didn't mention the putative CG15284 earlier as it only as a low Genie score.
This is the cystine-knot domain sequence (assume pu is CG15284) muc are the
human mucins
\* \* \* ^ \* \* \*\*
\* \* \* \*
This is the 3'terminal of the mucin proteins. The cys residues make up a
cystine-knot and are involved in the oligomerisation of the mucins. ^ is an
important conserved Glycine
Norries disease protein The first amino acids are a signal peptide
1 a 2 3
| : :||: : | |:: :||| | || |: |
b \*4 c 5 6 d
 :: | ||| | | : | :| | ||| |:| |
The conserved cys residues are numbered and form cys-cys bonds as follows:
 1:4 ,  2:5 ,  3:6 
The cys marked with \* is essential to cross link monomer units.
The degree of cross homology is similar to that seen between human TGF-b and
NDP, the spacing between the cys residues not too important.
Note that this is the same level of cross-homology as between other
cystine-knot domains
The same region is conserved in von Willebrand Factor
Yours sincerely
Terry Davis
Associated Information
Associated Files
Other Information
Secondary IDs
    Language of Publication
    Additional Languages of Abstract
    Parent Publication
    Publication Type
    Data From Reference
    Aberrations (9)
    Genes (3)