Heterozygous mutant cannot produce homozygous


Has anyone come across the situation where a heterozygous mutant doesn’t give homozygous progeny?

I tried to generate a splicing site mutation using CRISPR and 2 F1 progeny seemed to be heterozygous (after DraI digestion, wt should show 341+114bp and designed mutant should show 91+114+248bp). Then I picked 20 F2s from one of them and did SW PCR. However, no homozygous mutant in all the F2s. I further checked several F3s from one heterozygous F2 and still no homozygous.

I’m sure that this gene is not lethal and I have generated another very similar mutant to this design and got homozygous progeny from a heterozygous parent.


Thank you for your attention.

I haven’t come across this kind of situation. Did you pick the first a few grown-up F2s for genotyping? Or did you pick F2s generated overtime? That could make a difference if the homozygous causes developmental delay.
That’s all I can think of based on the info provided. Maybe others have better suggestions.

One obvious possibility would be a tandem duplication containing a wild-type and a mutant copy of the locus. You might be able to resolve this by balancing the locus, preferably with a gfp transgene or other dominant marker, to get animals you’re pretty sure have homozygosed the locus.

Or as Flora said, maybe you just can’t homozygose this allele because it’s dead or slow growing and so unrepresented, either because it behaves differently from other alleles or because from bad luck it is linked to something that does.

Hi Flora, thank you for your reply.
I tried late grown F2s and still no homozygous. I also randomly picked starved L1s and they showed almost 1:1 ratio of “wt” and heterozygous.
Actually, I sequenced the wt-like band and heterozygous band and found that the wt-like band has a small deletion that outside my target edition site. For the heterozygous band, since it contains 2 sequences with some overlap, based on the 2 peak signals in each position in part of the seq result, I can roughly tell that the other sequence is my target mutation. So the heterozygous contains 2 different mutations and it seems they cannot be separated.

You may want to run more positive / negative controls. e.g. N2, digested/undigested (after PCR), and then also try different enzymes. Your digestion doesn’t make sense.

The ‘bottom’ band runs at the same position whether WT or a het. If this is 114 bp where is the 91 bp band? does it exist?

If the top band in a het is 248 bp, then why does it run at the same position in your F1 gel in hets and WT? i.e. why can you not distinguish a 248 bp band from a 348 bp band about 50 bp apart?

If you can’t distinguish 2 bands 50 bp apart why should you expect to be able to tell 91bp and 114 bp bands apart?

It just doesn’t seem the digestion is working as expected.

Hi Richard, thanks for replying.

The last 2 lanes of the first gel is wt with and without digestion, and it gives 2 bands of 114+341bp. The top band of the het is 341 as in wt and the second band is 248bp. The 91 bp band may be very close to 114bp band or too dim to be seen. Nevertheless, the 248bp band suggests the mutation occurred, but it’s heterozygous with wt.
I cannot try another enzyme since the difference I introduced can only be detected by DraI. I also tried enzyme digestion after purification of pcr band and it showed the same pattern as digestion in pcr reaction mix.

I confirmed that the 3 band digestion pattern is heterozygous by sequencing(the previous reply) although it is a hetero of 2 different mutations.

Hi Hillel
Can you explain more about the “tandem duplication” and how can I separate them by balancing?
Thank you a lot.

A duplication would mean that the genome now had two copies of the locus (in this case, one mutated and one not) instead of just one as in the wild type. A tandem duplication would mean that the new copy was very close to the original location, right next to it. You could use a balancer to isolate that locus from the mutated strain, and to homozygose it. If a homozygosed copy of the locus from the mutated strain genotypes as having both wild-type and mutated copies of the locus, then you have a duplication, probably a duplication linked to the original locus, such as a tandem duplication. So, a balancer could inform you as to what had happened. Unfortunately, if this does turn out to be the case there is no good way to separate them.

The balancer is worth using in any case, as it’s a good way to figure out if homozygotes for the mutated locus are dead or otherwise underrepresented.

When these types of mystery problems arose in our screens, it was usually technical due to PCR/digestion. First thing I would do is to Sanger sequence the full PCR from one of your hets and make sure that this is in fact a het by looking at the traces.