288 L. G. AUGENSTINE 



formation of a single new bond (4)*, it is probably produced most often by 

 the spontaneous breaking of a large number of intramolecular bonds which 

 is accompanied by a very large increase in entropy. Steps 1 and 2 would 

 constitute the activated state of physical denaturation, i.e., reversible inacti- 

 vation, whereas the rupture of the second S — S bond, step 3, allows irreversible 

 inactivation to proceed. The large entropy change often found to be associated 

 with irreversible denaturation indicates that a partial unfolding of the molecule 

 usually occurs, and therefore the 'weak-link' is probably involved in latching 

 the molecule together. Once the second S — S bond has ruptured, the degree 

 of denaturation will depend both upon the extent to which unfolding proceeds 

 and upon subsequent reactions of the newly exposed groups of the altered 

 molecule. The number and the location of the intramolecular bonds involved 

 in step 2 is thought to be essentially invariable for a given protein, and to depend 

 upon its particular structure in the region of the weak link. It is postulated 

 that a variable number of bonds is involved in step 2 since the enthalpy of 

 denaturation activation varies widely between different proteins (8, 1). 



These generalizations are consistent with a variety of experimental findings, 

 many of which have been discussed elsewhere (1,2,3), and therefore will be 

 only mentioned briefly. 



(a) Data reported by a number of investigators give a value for the free 

 energy of denaturation activation of AF* = 24.8 i 1-5 kcal/mole (or 1.1 

 eV/molecule), which is slightly in excess of that required to rupture an S — S 

 bond (8). 



(b) Mild denaturation is reversible, whereas violent denaturation is not. 



(c) Following activation an additional 20 kcal/mole (for trypsin) is necessary 

 to initiate a large entropy change (about four or five times that for A^* of 

 activation). This is thought to be a large configurational change. 



(d) An average of two to three cysteine equivalents per insuhn molecule 

 corresponds to a fifty per cent reduction in its biological activity (the present 

 hypothesis would predict three cysteine equivalents per molecule, i.e. two 

 for each reversibly inactivated and four for each irreversibly inactivated mole- 

 cule). 



(e) The appearance of the full protein sulhydryl titer is invariably associated 

 with complete loss in activity. 



(f) Disulfide bonds are likely involved in the latching together of large 

 segments of the insulin molecule, since reoxidation of the reduced insulin 

 molecule causes aggregation. 



(g) The ultraviolet action spectra for the inactivation of trypsin and ribo- 

 nuclease, both of which have high cystine contents, are peaked at a wavelength 

 corresponding to maximum cystine absorption, and the quantum efficiency 

 is strongly correlated to their cystine content (9), (10); however, Setlow and 

 Doyle (10) found that the action spectra for gramicidin and aldolase, which 

 had little or no cystine, roughly paralleled the molecular absorption spectra. 

 They concluded that although there must be more than one inactivation 

 mechanism, a quantum absorbed by cystine could be as much as twenty-five 

 times as effective in producing inactivation as one absorbed by an arom.atic 



♦ For instance, inactivation due to freezing and drying, which is apparently not accom- 

 panied by a gross opening of the molecule (5), may depend upon such a process. 



