How did you obtain the cDNAs for the LC and HC? Could there be a codon usage problem (low frequency?) and cryptic introns/splice sites?

Wolfgang Schechinger

Thanks for the comment.

The sequences were from an established IgG HC/LC amino acid sequence, codon-optimized for CHO cells on Genscript, and synthesized/sequence verified. I checked codon usage on bioinformatic.org, the frequency corresponded to CHO cells.

For splicing, I am not an expert. But I used https://www.fruitfly.org/seq_tools/splice.html to predict both sequences, there are some (

To acquire the best yield of protein, maybe try to express Heavy chain and light chain independently with various ratio can help.

also

1. if you use antibody to recognize region of LC+HC, I am curious how can you detect the signal of 25kDa (LC only)?

2. I think the interaction between protein A and LC can not illustrate your high yield of LC. Since if the interaction between protein A and LC are weak, LC can not even bind on it which means there will be no or low LC in your elution fraction, and the signal should be shown in the flowthrough fraction.

3. I suggest to clone the variable region in the vector with the backbone of constant region. This way should be easier, and the result should be predictable. Besides, maybe try to add small tag (ex: his or flag) on LC or HC can help you to troubleshoot in the beginning.

4 The yield in conditioned media is more important .I think it will be fine if there are a great number of LC+HC antibodies in the conditioned media.

Good luck

Lo Tzu Yu

Thanks for the suggestions. We are in fact trying to use a higher HC/LC ratio (3:2, 4:1) for expression, also in an alternative host. I'll update if it worked.

Regarding your question of LC/HC, the secondary antibody is anti-both chains (goat anti-human) We frequently use this antibody for IgG intracellular applications and are able to see both chains.

The current problem is that in conditioned media, there's also a majority of light chains and the assembled IgG is not abundant, seen in non-reducing CE-SDS. We're staining intracellular bands to rule out some contamination effects (bound IgG structures on the protein A column not eluted fully from the previous trials). But as you mentioned, it's weird that LC itself should not bind very strongly to proA. So we're also working on troubleshooting the downstream purification to make sure there's no strongly reducing events - one possible event might be shear during concentation? We frequently use a 10kDa MWCO spin column at 14,000 xg. Our IgG from a stable cell line was very tolerant but maybe not those transiently transfected ones. If our current trial did not work, we'll certainly think more about where to proceed from your suggestions. Thanks again.

Leran Mao

I think maybe it is better to check the size of your LC, HC in conditioned media via WB before purification .This way should help you to distinguish whether the problem is from expression or purification. (though I think you actually done it ?)

please update your progress, I think it is a really interesting question

thanks for your sharing

good luck