Interesting observations. One of my colleagues recently got some weird results with their LR clones, too, so this may be common. At least you aren't alone!
When you say that you see your donor vector, do you mean the GOI-bearing entry vector? With Primer #3 do you see attL or attP site as well as your attB1 fragment and GEO? The donor vector is regenerated in the LR reaction, but it should have different antibiotic resistance than the destination vector (like the Entry vector). Very odd indeed. I don't know if it's possible to pick up and re-purify the donor/entry vectors from the plate that receives your transformation. Seems highly unlikely - this is usually only a problem with colony PCR. That makes me think you do have a hybrid of some sort.
If I were in this position and suspected multiple plasmids, I would try a few things:
1) Transform the sequenced product into new cells and re-prep a few more colonies to see if the pattern repeats in another sequencing round. This should dilute out contaminant plasmids.
2) Use a restriction digest to screen your sample. If possible, select a unique cut site that is in the region of your GOI covered by Primer #3. Two products vs. a linearized plasmid might indicate a hybrid plasmid or multiple plasmids in your prep. Also try digests with enzymes that only recognize your entry and destination vectors, respectively.
3) Start over. Verify your entry and destination sequences, get super clean DNA and perform the LR reaction again. The amount of DNA can be varied in the reaction according to the LR clonase manual - so I might play around with the amount of each vector. Definitely consider the stoichiometry of your entry and destination vectors (i.e. how many molecules do you have of each per given mass of DNA).
If you have multiple plasmids, I can only surmise that you've picked one up from a contaminating source in the lab (plasmids get everywhere). Re-transforming and carefully prepping your colonies should clean it up. If this were the case though, I'd also suspect that your sequencing traces would show dirty base calls from primer #3, indicative of two different, equal-length termination products rooted with the same primer sequence. This would begin approximately where the GOI sequence ends as that region should be common to any plasmid with the GOI insert.
Yes, by the donor vector I did mean the entry vector with my gene of interest
With primer 3 I see my gene of interest (where the primer binds) followed by a small part of the Attb1 site, followed by the entry vector (I guess a piece of the attL site).
So it looks like an incomplete LR reaction somehow.
But its weird I also (with primers 1 and 2) I get the correct expression clone.
And yes: it has a different antibiotic selection, so its weird I do seem to find it.
I also find it, in general, weird I would have 2 plasmids in 1 E.coli? Or maybe 1 hybrid plasmid that contains the gene twice? But this seems weird, I should see this in a gel, that my vector is too large for what it should be!
1) I did this and it seemed to have done the trick for 2 plasmids!
2) Working on this one.
3) well... its pretty expensive... I did it 4 times already by now and I changed pretty much everything I could change...
About the sequencing: you are right, when the donor/entry vector pops up in the sequencing results, it seems to be that there are 2 plasmids present...
After resequencing some plasmids that I transformed again, it gave me the correct sequence (the expression vector) and it seemed ok in terms of sequencing (nice peaks).
Biut this only worked for 3 plasmids out of 10 or something.