That depends what you are trying to do:

1- if trying to express in e.coli for example and the gene contains introns then use cDNA to get the gene without the introns.

2- if intronless gene you can use gDNA

in the end it depends on your target vector.

I'm preparing the plasmids for an overexpression assay. I'll transform in E.coli then collect the plasmids for transfection in a human cell line.

A primary distinction to be made between cDNA and gDNA is in the existence of introns and exons. ... cDNA also does not contain any other gDNA that does not directly code for a protein (referred to as non coding DNA). Lastly, not all genes in the gDNA are being transcribed into mRNA at any given time Abiola Afonja

Thanks Hanna and Shivam for the responses. My questions was do I isolate gDNA and insert this into my plasmid vector for cloning or do I isolate RNA, reverse transcribe and then insert the resulting cDNA into the plasmid vector.

As I said above. It depends on your final experiment. In your case it is over expression in mammalian cells. I am assuming you want as high expression as you can get hence I would suggest to check your particular gene, if the presence of the native introns enhances the expression then use gDNA, if not use cDNA. Also check if there are isoforms from alternative splicing. In case you need a specific isoform then use cDNA.