Stress sensors are specific detectors, not a general alarm system. Each one is coupled to a particular disturbance, and the identity of the sensor determines which downstream program is recruited.
Misfolded or unfolded protein in the cytosol is detected indirectly. Heat shock factor 1 (HSF1) is a transcription factor that is normally kept monomeric and inactive by binding to heat shock proteins (HSPs), especially Hsp70 and Hsp90. When misfolded protein accumulates, it sequesters these chaperones, freeing HSF1 to trimerize, enter the nucleus, and activate transcription. The sensor here is effectively the chaperone pool itself: HSF1 measures free chaperone availability, not the misfolded protein directly.
Amino acid limitation and other stresses that stall translation are detected by kinases that phosphorylate the alpha subunit of eukaryotic initiation factor 2 (eIF2alpha). Four kinases feed this node: GCN2 responds to uncharged tRNA during amino acid starvation, PKR responds to double-stranded RNA during viral infection, PERK responds to unfolded protein in the endoplasmic reticulum, and HRI responds to heme deficiency and oxidative stress in erythroid cells. Converging on one phosphorylation site is what makes this the integrated stress response.
Energy charge is detected by AMP-activated protein kinase (AMPK), which senses a rising AMP to ATP ratio. DNA damage is detected by ATM and ATR, which recognize breaks and replication stress and phosphorylate the checkpoint kinases CHK1 and CHK2. Oxidative stress is detected through oxidation of specific cysteine residues in sensor proteins such as KEAP1, which releases NRF2 to drive antioxidant gene expression.
The pattern to notice is that sensors are chemically matched to their input: a kinase for energy charge, a tRNA-binding kinase for amino acid supply, a chaperone-titration mechanism for folding load. This is why different stresses produce different, not interchangeable, responses.