If you’re interested in:

• gene regulatory network evolution
• transcription factor promiscuity
• predictability of adaptation
• or why evolution so often repeats itself

give the paper a read and let me know your thoughts.

The same principles likely extend beyond bacteria.

Across life, regulatory novelty often arises by redeploying existing network components - not by inventing new ones.

Evolution tinkers with what is already there.

These findings help reframe crosstalk.

Rather than asking
“why is crosstalk tolerated?”

we should ask:
how does crosstalk shape evolvability?

Importantly, TF recruitment is not well predicted by:

✗ sequence similarity alone
✗ structural similarity alone
✗ DNA-binding domain similarity alone

Context matters more than identity.

In our system, evolution often proceeds in two steps:

1. mutations increase TF activity/expression
2. subsequent mutations refine binding towards new promoters

And which TF is rewired is predictable based on the path of least resistance: the most mutationally accessible route.

That third point is key.

Low-level, non-cognate binding (aka crosstalk) provides the raw material for innovation.

Selection doesn’t invent new interactions from scratch.
It amplifies ones that already exist weakly.

Why these TFs and not others?

Our work shows three properties matter most:

• ability to reach high activity
• ability to reach high expression
• some pre-existing low-level affinity for new targets

Only a small subset of transcription factors can take over the lost function.
And they do so in a clear hierarchy of accessibility:
some regulators are evolutionary “first responders”.

When the flagellar master regulator (FleQ) is deleted, bacteria repeatedly evolve motility back - but not randomly.
They follow highly predictable evolutionary routes.

Using experimental evolution in Pseudomonas fluorescens, we can watch regulatory rewiring happen in real time.

Delete a master regulator → apply selection → observe evolutionary rescue.

This “promiscuous” binding is usually viewed as noise and costly misregulation that cells have evolved to suppress (and that is true for locally adapted organisms!).

But what if those mistakes can also matter for evolution? And is it possible to define it's role?

Gene regulation is often described as neat and deterministic, with each transcription factor (TF) controlling its own specific targets.
Reality is messier: They make mistakes; They engage in crosstalk/promiscuous binding

🙋🏻‍♀️

@yueyicai.bsky.social