I ran across the same problem and found that Clontech (now TAKRA I think) has a version they call IRES2 that isn't attenuated.  I used that when I wanted both genes to express at high levels.

Dear Cristoph

You should check out DNA 2.0: https://www.dna20.com/products/expression-vectors/mammalian. They have a range of vectors available that should suit your needs. you could also consider to use the self cleavable t2A peptide instead of an ires as it will aid in strong expression of both genes.

best

Kai

Why don't you have it synthesized exactly the way you want...  That might even be cheaper than whatever people think they can charge for plasmids, and saves you the trouble of subcloning. 

If you want achieve equal levels of expression for two genes, better use 2A peptide and you don't need any special vector for this, just order the primers. If you really want an IRES vector you can PCR amplify IRES from MIGR (available at addgene I think) and PCR ligate it with your two cDNAs. I have MIGR plasmid, it's CMV-IRES-GFP, and this IRES gives 8 times more GFP than the 1st gene before IRES. Or you can synthesize it, just as suggested above by John S, flank it with oligos matching your target cDNAs and then PCR the whole thing, it´s the fastest way, if your cDNAs are not too big (IRES is around 600 bp)

best

Geni

I have created this kind  of vector by inserting the IRES and two MCS into pCDNA5FRT-to vector. If you want I can send it to you. Just contact me by mail or tell me your email address

However the problem of lower expression of the second cassette exists and it is a common problem  of IRES containing vectors., it is correct what was stated in the post of Evguenia Bekman, if you require the same expression level of the two genes  it is better to use he 2A peptide , you can also express 3-4 genes simultaneously

all the best

Anna

The pCMMP-MCS-IRES-mRFP vector from addgene is good.  Since the mRFP is behind the IRES, as long as you see red fluorescence it's a good indication you're getting expression of the gene you put in the MCS as well.

Thanks for your answers! 

Indeed it's right, that clontech is also offering plasmids with a non-attenuated IRES, called IRES2. I think I'll use them in the future. 

The problem with the P2A sites is, that my "two proteins" are acutally already four. Two of them are fused with P2A sites. I was just concerned about too many N/C- terminal modifications and also I'm not sure about how many proteins you can really express with comparable expression levels with P2A. Any experience on this?

I know that three or four different proteins can be efficiently expressed using different 2As. Some proteins are more, others, less sensitive to the addition of N/C terminal tags, unfortunately, until you try with your protein, you  can never be 100% sure it will work. So, good luck!

i think its better to construct ur own

I just compared the IRES sequences of pIRES and pIRES2. They only differ in that IRES2 has a few additional bases in the end: tatggccacaacc

Is that what they mean by "a partially disabled IRES sequence" ?

Dear Clemens,

I guess it is M-loop at 3' end of EMCV IRES. Check Bochkov and Palmenberg 2006 BioTechniques paper (https://www.future-science.com/doi/10.2144/000112243?url_ver=Z39.88-2003&rfr_id=ori:rid:crossref.org&rfr_dat=cr_pub%3dwww.ncbi.nlm.nih.gov), they have a nice IRES map there and also discuss historical development of commercial IRESs, including pIRES. They also define minimum and preferred IRES sequences, with M-loop in the preferred part.

In Chamond et al. 2014 Nucleic Acids Research paper (Article 40S recruitment in the absence of eIF4G/4A by EMCV IRES refi...

Thanks for your comment, it's quite useful for me!

I can tell you from personal experience that the IRES is attenuated and the IRES2 gives full expression. I had it in front of the EGFP gene and I couldn't see it under the microscope until I used the IRES2 version.

The reason many people use the attenuated version is to avoid sometimes toxic over expression so it depends on your needs. As mentioned above, however, this doesn't mean the expression levels will be identical even if they are on one transcript.