For the
second delay, consider sequential independent
first-order transitions
A⟶kaB⟶kbC. Their dwell
times have
exponential distributions. The total waiting
time is their
sum, so
convolution gives
P2(t)=∫0tkae−kaskbe−kb(t−s)ds=ka−kbkakb(e−kbt−e−kat),t≥0.
It is normalized, symmetric under exchange of
ka,kb, and has
⟨t2⟩=ka1+kb1,Var(t2)=ka21+kb21.
Unlike one exponential delay, it vanishes at
t=0, rises initially
as kakbt, reaches
a maximum at
tmode=log(ka/kb)/(ka−kb) and eventually decays at the slower rate. For equal
rates the continuous limit is
P2(t)=k2te−kt, an
Erlang distribution with mode
1/k. If one rate is much faster than the other, the observed long-
time behavior approaches one exponential. The sketch in the preceding section displays these
shapes using arbitrary specified
rates.
The simplest
catalytic cycle compatible with the fits therefore contains an ATP-dependent binding delay and at least two resolvable ATP-
concentration-independent kinetic
events.
Hydrolysis, product release and conformational changes are possible microscopic assignments, but the two-rate fit alone does not uniquely identify them or the order of
ka and
kb. If each complete
120∘ cycle contains these three waits and step durations are negligible, its
mean duration is
(k1c)−1+ka−1+kb−1 and the
mean revolution rate is the inverse of three
times this duration. Multiple catalytic sites can cooperate during that cycle.
TODO: the source PDF's stepping plot is absent. Its individual substep angles, numerical timings and the mapping of plotted dwell labels onto the larger or smaller angular movement cannot be read. The surviving text establishes only the
120∘ total step and the
existence of
a larger and
a smaller substep. No numerical rate, missing
angle or plotted label is inferred from
a different
experiment. The distribution derivations and the
concentration-dependence
inference above are independent of that
illustration.