Summary Probably ApoE E2/E4, but maybe E1/E3. E1 is the 'missing ApoE allele'
Numerous people have now confirmed seeing this with no apparent consequences. It is interesting, but doesn't seem worrisome.
[PMID 17244188] This genoset looks for the possibility of ApoE1 'the missing APOE allele'. It is unclear why ApoE1 is rare. You have rs429358(C) and rs7412(T), but we cannot tell if they are in the same phase.
However we've still never had a gs267 report in. So it seems increasingly likely that you're E2/E4 and that E1 is really really rare.
The 'missing allele' ApoE1 is visible on the top strand. The bottom strand shows an ApoE3
The top strand is an ApoE2. The bottom strand is an ApoE4
This reference says that the rare APOE has been seen a few times before. In the paper, the person who had the rare APOE variant had ALS, Lou Gehrig's disease! Half of the others with the variant did not seem to have a severe disorder, while the others did. But they are talking about very small numbers.
There is also large bias, because healthy people weren't being sequenced much in 2007. The 4 people who've seen gs189 in their Promethease are completely healthy. Even if you're not healthy, it's unlikely that gs189 contributes to what ails you. gs189 highlights a mystery of all human genomes. Hopefully soon we'll learn more about this, and you may be one of the ones who can help us to answer the question. Whether or not you have it, there does appear to be a rare genotype, and you're one of the small number of people who can provide unique data. As you learn more please update this page.
I'm assuming I would need a more advanced genome scan to find out, that is, if there are not nice SNP's already measured on the 23andme chip that would help determine if I am E2/E4 like 23andme says, or E2/E3/unnamed?
It is possible other nearby snps could help to indicate the phasing of your data, but currently we've no further information.
How would I find out if I have the unusual variant?
You only want to know the full sequence of one gene, so you don't need a full genome scan. This type of work would cost less than $1000 to perform in a traditional lab, often MUCH less. While they probably exist, I'm uninformed on labs which are sufficiently 'consumer oriented' for your needs. If you're inclined, Yahoo answers has information on sequencing yourself, and perhaps the DIYbio community would be able to provide more information and resources.
There is also some discussion in this 23andMe forum post
Previously also known as gs143 the duplicate names were accidental and have been consolidated