DNA in viruses and cells exists in highly condensed, tightly packaged state
s. We have undertaken an in vitro study of the kinetics of DNA condensation
by the trivalent cation hexaammine cobalt (III) with the aim of formulatin
g a quantitative, mechanistic model of the condensation process. Experiment
al approaches included total intensity and dynamic light scattering, electr
on microscopy, and differential sedimentation. We determined the average de
gree of condensation, the distribution of condensate sizes, and the fractio
n of uncondensed DNA as a function of reaction time for a range of [DNA] an
d [Co(NH3)(6)(3+)]. We find the following: (1) DNA condensation occurs only
above a critical [Co(NH3)(6)(3+)] for a given DNA and salt concentration.
At the onset of condensation, [Co(NH3)(6)(3+)]/[DNA-phosphate] is close to
the average value of 0.54, which reflects the 89-90% charge neutralization
criterion for condensation. (2) The equilibrium weight average hydrodynamic
radius [R-H] of the condensates first decreases, then increases vith incre
asing [Co(NH3)(6)(3+)] as they undergo a transition from intramolecular (mo
nomolecular): to intermolecular (multimolecular) condensation, However, [R-
H] is insensitive to [DNA]. (3) The uncondensed DNA fraction decays approxi
mately exponentially with rime. The equilibrium uncondensed DNA fraction an
d relaxation time decrease with increasing [Co(NH3)(6)(3+)] but are insensi
tive to [DNA]. (4) The condensation rate in its early stages is insensitive
to [DNA] but proportional to [Co(NH3)(6)(3+)](xs) = [Co(NH3)(6)(3+)] - [Co
(NH3)(6)(3+)](crit). (5) Data for low [DNA] and low [Co(NH3)(6)(3+)] at ear
ly stages of condensation are most reliable for kinetic modeling since unde
r these conditions there is minimal clumping and network formation among se
parate condensates. A mechanism with initial monomolecular nucleation and s
ubsequent bimolecular association and unimolecular dissociation steps with
rate constants that depend on the number of DNA molecules in the condensate
, accounts reasonably well for these observations. (C) 2000 John Wiley & So
ns, Inc.