Worm Breeder's Gazette 11(4): 90

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.

A Genome-Wide Screen for Zygotic Embryonic Lethal Mutants

Joel H. Rothman

Figure 1

One way to identify essential genes likely to play controlling roles 
in embryonic development is to screen for zygotic embryonic lethal (
ZEL) mutations covered by a duplication or balancer.  However, such 
screens limit one's search to a fraction of the genome, thereby 
preventing an assessment of a broad range of phenotypes.  With the aim 
of analyzing a large proportion of all possible ZEL mutants in the 
worm, I have developed a procedure for screening the entire genome 
without the limitations imposed by balancers.
In the screen, F1 self-progeny of mutagenized hermaphrodites are 
singled into microtiter wells and removed after 2 days of egg-laying.  
The F2 brood is then screened for ~25% arrested embryos 12 hours later.
The screen is repeated on 6 F2 surviving siblings from each of the 
wells that contained 25% arrested embryos.  Rescreened F2's that 
produce 25% dead eggs in the expected ratios are allowed to propagate 
to F4 broods of about 10,000 starved L1's, which are immediately 
frozen on the 35mm plates used in the follow-up screen.  Up to 40% of 
the putative ZEL's identified in the primary screen behave aberrantly 
in the subsequent screen and are discarded, demonstrating the 
importance of the secondary screen.  Initial attempts at this 
unbalanced screen were complicated by the finding that at standard 
levels of mutagenesis most wells contained some dead eggs (comprising 
from 1 to 100% of the F2's), and a high proportion of the 25% class 
failed to show transmission characteristic of a single locus ZEL.  I 
found that by lowering the mutagenesis level the aberrant classes 
disappeared at a much greater rate than the ~25% class; at an 
optimized level, most of the mutants passing both screens are bona 
fide ZEL's.
This screening procedure has allowed me to build up an extensive 
permanent collection of ZEL's.  In each frozen plate there are enough 
worms to allow repeated sampling (~10 times) of many F4's without 
having to reisolate the lethal mutation.  ~40% of the frozen worms are 
heterozygotes, providing sufficient fresh homozygous F5 embryos for 
screening by Nomarski, or with antibodies.  To date, I've isolated 572 
independent ZEL's from 15,800 F1's.  I would like ultimately to 
approach saturation of the genome.  Depending on the method of 
calculation, my current collection should have already identified on 
the average ~1-2 alleles per genomic ZEL locus.
For the primary phenotypic screening, I've been examining the 
terminal phenotypes of all mutants by Nomarski.  So far I've looked at 
405 ZEL's and have carried out mapping, staining with antibodies, and 
lineage analysis on a selection of those with the most intriguing 
phenotypes.  11 ZEL's, selected on the basis of their interesting 
terminal phenotypes, mapped to 5 chromosomes as follows: LGI -- 2 
mutants; LGII -- 4 mutants; LGIV --1 mutant; LGV -- 3 mutants; LGX --1 
mutant.  Therefore, as expected, the screening procedure works to 
identify ZEL genes broadly distributed over the entire genome.  One 
mutation of interest mapped far out on the right arm of LGV -- a 
region that may never be covered by a balancer chromosome owing to its 
proximity to the presumptive pairing site; this illustrates one 
advantage of this screen over a balanced lethal screen.  There are two 
cases in which pairs of mutants, selected on the basis of their 
similar unique phenotypes, mapped to similar regions (and are probably 
allelic), demonstrating that it is possible to identify multiple 
alleles of unique genes reproducibly based solely on their terminal 
The distribution of general arrest phenotypes for all 405 mutants is 
shown in the table.  The most common arrest phenotype is a pretzel 
with no obvious defects; most such mutants are probably defective in 
'housekeeping' genes.  None of the mutants arrests uniformly with 
substantially less than the normal complement of cells (500) in mature 
wild-type embryos.
The phenotypes of some of the mutants are particularly striking.  
One mutant (e2483 X) shows a (partial) block of gut differentiation, 
and arrests as an irregularly shaped, partially elongated embryo.  Six 
mutants [e2503 II; ZE199 & ZE312, LGI (probably allelic); ZE200, LGIV; 
ZE372; ZE378] generally fail to undergo pharynx morphogenesis 
similarly to pha-1 mutants (Schnabel and Schnabel, Science, in press), 
but also show pleiotropic effects on elongation, arresting at some 
stage between the onset of elongation (i.e., the hypodermis encloses 
the embryo, but elongation hasn't begun) and the pretzel stage.  These 
mutants generally lack a pharynx lumen, grinder, buccal cavity, and 
organized muscle fibers, but do have a basement membrane surrounding 
an under-developed pharynx.  In several [ZE199 (and probably ZE312), 
ZE200, ZE372, and ZE378], expression of late pharynx antigens, such as 
myosin C and a gland-specific antigen are blocked; earlier antigens 
are generally unaffected.  The defects in pharynx differentiation are 
not likely to be the result of a failure in elongation per se, since 
some mutants that don't initiate elongation nevertheless express late 
pharynx antigens (see accompanying article).  Two allelic mutations (
e2501 II and ZE256, LGII) result in arrest as irregularly shaped 
partially elongated embryos with an abnormally large number (~15) of 
persistent cell deaths in the head, an incompletely formed pharynx and 
generally no discernible rectum.  Finally, only five mutants arrested 
uniformly prior to the onset of elongation, resulting in embryos that 
lack hypodermis over most of their surfaces; this class of mutants is 
discussed more extensively in the accompanying article.
Continued screening and in-depth analysis will allow me to assess 
more completely the range of developmental events during embryogenesis 
that are under control of the zygotic genome.  A particularly 
interesting class of genes to me are those that control hypodermal 
fates and behavior, I believe many of these are likely to be 
identified among the classes of mutants that arrest a) prior to 
elongation (5 mutants), b) as partially elongated embryos with 
irregular shapes (17 mutants), and c) with loss of hypodermal 
integrity during elongation (32 mutants).  Analysis of these mutants 
with antibodies to hypodermal antigens, and by lineaging, should allow 
me to identify genes required for specification and function of the 
[See Figure 1]

Figure 1