You should indeed get rid of the IRES as well to get a "normal" Puro expression construct. A Kozak sequence at the Puro start ATG is recommended. However if you just want to have a CMV-Puro construct, I would rather go for an established construct to be on the safe side. (ask a colleague or try adgene.org).

Uwe Borgmeyer Thank you very much for your answer!

Would you be so kind to explain in what way the presence of IRES without GFP could disturb the expression of PuroR?

Or maybe you can advise me some literature where I can learn about that? I was struggling to find the literature myself...

Addgene has a lot of free information. See Plasmid 101 book and

https://blog.addgene.org/plasmids-101-multicistronic-vectors

Uwe Borgmeyer , thank you for the link. However, I found there only general information about the mutlicistronic vectors. I do not have problems with finding such information, but I have problems with finding information on how exactly the lack of Gene1 (GFP in my case) between promoter and IRES could influence the expression of Gene2 (Puro in my case).

Would you be able to send me some sources on this precise question?

Without deletion, you have single long RNA which codes for two proteins. Gene 2 translation is based on the IRES. The expression of gene 2 is in many cases lower (see link above). If you get rid of GFP and the IRES you have a classical transcript driven by the CMV promoter. The construct is smaller and the transcript shorter. This may lead to higher expression. With a Kozak sequence at the start ATG sequence you can expect high expression of Puro.

However, you can not predict the exact outcome.

Uwe Borgmeyer Aha, I see. Yes, I've read about the Gene2 expression being lower.

So, as I understand that is the only drastic effect with cutting the Gene1 or maybe you are aware of some additional effects?

Thank you

There should be no additional effects.