Leaky protein expression in your pET system is likely due to a combination of factors, and based on your observation that previously stored GFP plasmids showed no leakage while recent transformations do, the most probable causes involve changes in host cell strain physiology, repressor function, or plasmid integrity. The pET system relies on the T7 promoter, which is tightly controlled by the lac operon, particularly via the lacI repressor and host strain compatibility. If your competent E. coli strain (e.g., BL21(DE3)) has compromised lacI function (due to mutations, plasmid loss, or degradation), basal transcription from the T7 promoter can occur even without IPTG, causing leaky expression. Additionally, overgrowth or stress during competent cell preparation can lead to unstable DE3 lysogen maintenance, reducing repression efficiency. Also, if the plasmid has undergone mutations, particularly in the T7 promoter region or lac operator site, or if plasmid copy number regulation has been altered (e.g., through loss of rop gene or ori mutations), promoter repression may weaken, increasing basal transcription. Another contributing factor is the presence of lactose or similar sugars in rich media like LB, which can cause partial induction even without IPTG. Lastly, if T7 RNA polymerase is overexpressed due to promoter leakiness or chromosomal derepression, even small amounts can drive strong expression from pET vectors. To resolve this, consider retransforming freshly prepared plasmid DNA into a verified, tightly regulated host strain (like BL21(DE3)pLysS), sequencing your plasmids to check for mutations, and using media that limits unintended induction.

Some questions first:

1. Are you comparing the fluorescence on a plate to the same strain transformed with an empty vector? If not, I would do that first to make sure it isn't just misidentified autofluorescence.

2. What E. coli strain are you using for transformation and is the strain you're using now different from the one that was used previously? For example, in BL21(DE3) expression of the T7 polymerase is under control of a stronger, but more leaky mutant variant of the lac promoter and tends to exhibit high basal expression vs. BL21(DE3)pLysS where basal amounts of T7 polymerase are sequestered by the T7 lysozyme. If you are using a strain that carries a plasmid to more tightly control expression, make sure the antibiotic for the control plasmid is included in the growth media when making your competent cells, otherwise some cells may randomly lose it.

Finally, I would look out for contamination with some type of species that isn't E. coli. For example, Pseudomonas species are pretty ubiquitous in the environment, can produce fluorescent pigments, and are often inherently resistant to a lot of common antibiotics like ampicillin and kanamycin.

Sometimes the culture media for making competent cells or the elution buffers for plasmid preps get contaminated with Pseudomonas or some other tough bacteria and they end up growing on the recovery plates. Fortunately, you should be able to rule this out pretty quickly by mock-transforming your competent cells without a plasmid and growing them on antibiotic selection plates. You shouldn't see any growth, if you do then either your competent cells are contaminated or your plates are going bad. You can also check for contamination in your plasmid prep or any buffers you're using just by spotting some of it on the antibiotic selection plate and streaking it out.