Worm Breeder's Gazette 11(4): 59

These abstracts should not be cited in bibliographies. Material contained herein should be treated as personal communication and should be cited as such only with the consent of the author.

Inheritance of Pattern? sqt-1 and rol-6 are Coordinately Expressed, but Not at All Stages that Roll

Jim Kramer and Yang-Seo Park

Figure 1

Coordinate expression.  We have examined expression of the sqt-1 and 
rol-6 collagen genes by Northern and slot blot analyses.  Both genes 
are expressed at the L1, L2, L3, and L4 molts, but no transcripts are 
detectable in mixed-stage embryo RNA or L2d-dauer RNA.  By 
quantitative slot blot analysis, the relative levels of sqt-1:rol-6 
mRNAs are approximately 2:1 at each of the molts at which they are 
expressed.  These results are consistent with the proposal, based on 
genetic interactions and structural similarity, that sqt-1 and rol-6 
physically interact and suggest that they could be present in a 2:1 
ratio within a single heterotrimeric collagen molecule.  The rol-6 
mRNA is trans-spliced with SL1, while sqt-1 has a cis-spliced intron 
in its 5'UTR; another example in which one of a pair of related 
collagen genes has switched its mode of splicing (the other is col-
12/13).  If SL1 affects translation, then the sqt-1:rol-6 polypeptide 
ratio may not be 2:1, but would still be constant at all stages.
Expression at the L2 stage.  sqt-1 and rol-6 transcripts are 
produced during the L1-L2 molt at levels similar to those seen in 
later molts, but mutant animals had been reported not to show 
phenotypes at the L2 stage (Cox et al.  Genetics 95:317, 1980).  We 
find that at the L2 stage e1350/+, e187 and sv1006 animals are strong 
RRol; some e1350 animals have slightly abnormal tails and move 
abnormally; sc13 animals are indistinguishable from N2.  Thus, sqt-1 
and rol-6 expression at the L1 molt can produce the RRol phenotype, 
but apparently not the LRol phenotype.  Possibly some components of 
the cuticle required to produce the LRol phenotype are not expressed 
until the L2-L3 molt.
The dilemma of dauers sqt-1 and rol-6 mutant animals display the Rol 
phenotype as dauers, but we could not detect transcripts from either 
gene in two different preparations of L2d-dauer RNA.  To substantiate 
these results, we isolated RNA from daf-2(e1368) animals at multiple 
time points throughout the L1-L2d-dauer period and probed slot blots 
with sqt-1, -specific collagen) 
specific probes.  sqt-1 and rol-6 transcripts were detected at and 
immediately following the L1-L2d molt (16-22 hours after plating L1s 
at 25 C), but were undetectable at all later time points (26-54 hrs).  
In contrast, col-2 transcripts were first detected when animals began 
the L2d-dauer molt (30 hrs), peaked at 38 hrs., and were barely 
detectable when animals had completed the molt (54 hrs).  Given the 
level of sensitivity of these experiments, the levels of the sqt-1 and 
rol-6 transcripts must be at least 50-100 fold lower at the L2d-dauer 
molt than at other molts.
We examined sqt-1 and rol-6 mutant animals at the L2d and dauer 
stages.  Similar to L2 stage animals, e1350/+, e187 and su1006 animals 
are strong RRol as dauers.  sc13 and e1350 dauer animals display 
variable LRol or RRol phenotypes, respectively.  The term variable is 
used to indicate that some animals show strong Rol phenotype, some are 
weak Rol, and some are not observed to roll.  The L2d phenotypes for 
these animals are the same as their dauer phenotypes.  Note that the 
sqt-1 alleles display different phenotypes at the L2 and L2d stages, 
directly demonstrating a functional difference between these cuticles.
Inheritance of pattern?  How can the apparent lack of gene 
expression during formation of the dauer cuticle be reconciled with 
the appearance of mutant phenotypes at that stage (recall that the 
null phenotypes for both sqt-1 and rol-6 are WT)? It is possible that 
there is a very low level of sqt-1 and rol-6 expression and that it is 
sufficient to cause a phenotype.  This seems an unlikely explanation 
for these structural proteins, given that the expression is at least 
50-fold less than at an equivalent stage (L2-L3).  It is possible that 
sqt-1 and rol-6 collagens could be extracted from the L2d cuticle and 
reutilized for assembly of the dauer cuticle.  The strongest argument 
against this possibility is that the collagens in the cuticle are 
crosslinked with di- and trityrosine residues (D.  Eyre and J.  Kramer,
unpublished results).  There is no know mechanism to sever these 
crosslinks, short of hydrolyzing the proteins.  Another possibility is 
that dauers are rollers because the preceding L2d stage animals were 
rollers, i.e., the pattern is maintained from one stage to the next in 
the absence of expression of the mutant collagen.  We favor this 
explanation based on the arguments that follow.  At each molt a new 
cuticle is assembled beneath the old cuticle, so connections of the 
hypodermis and muscles to the cuticle must be broken during the molt.  
It seems likely that some sort of scaffold must form at this time to 
maintain organization in the absence of these cuticle attachments.  
Jim Priess showed that elongation of the embryo is dependent on 
bundles of microfilaments and microtubules circumferentially oriented 
in the hypodermis and that these filaments also appear to be involved 
in formation of the L1 cuticle (Priess and Hirsh Dev.  Biol.  117:156,
1986).  When the L1 cuticle has formed the filaments disappear.  He 
has found that at each molt the actin filaments reform, and could act 
as the postulated scaffold.  When we stained sc13 and e187 animals 
with TRITC-phalloidin, the hypodermal syncitia and seam cells were 
seen to remain helically twisted throughout the entire L2d-dauer 
molting period.  Bundles of circumferential actin filaments were seen 
in the hypodermal syncitia, while seam cells stained diffusely.  The 
filaments are oriented perpendicular to the seam-syncitium junction 
and appear to attach to the belt desmosomes.  This cytoskeletal 
arrangement presumably prevents the hypodermis from changing its 
organization, and locks in the helical structure present before the 
molt began.  Since the dauer cuticle is synthesized by a helically 
twisted hypodermis it may not require mutant collagen to generate its 
twist.  Thus, dauers inherit the pattern that existed in the L2d stage 
This proposal predicts that the dauer phenotype must be the same as 
the L2d phenotype, and this is true for the sqt-1 and rol-6 mutants we 
have examined.  It may also explain why, for certain morphological 
mutants, adults derived from dauers have a different phenotype than 
adults derived from L3s.  The persistent dumpiness induced by 
levamisole treatment (Lewis et al.  95:905,1980) could also be 
explained as maintenance of the hypercontracted morphology by the 
hypodermal filament system.  If this proposal is correct, the 
phenotypes of some morphological mutants could be a result of the 
combined effects of the morphology at previous stages and the 
composition of the current cuticle.
[See Figure 1]

Figure 1