purification, labeling of the molecule of interest with a suitable
marker and its re-introduction into the living system; (2) the use
of relatively large, bulky labels that can induce cross-linking of
target molecules or steric hindrance effects, (3) the need for a
large number of single-molecule trajectories to obtain trustable
statistics.
In this regard, fluorescence correlation spectroscopy (FCS) is
rapidly emerging as a very attractive experimental platform. In fact,
thanks to its intrinsic single-molecule “sensitivity” in the presence
of many similarly labeled molecules, it can easily afford the required
statistics in a limited amount of time. In addition, FCS works well
with genetically encoded fluorescent proteins and, in general, with
relatively dim fluorophores. The basic principle of fluctuation anal-
ysis is that the fluorescent molecules stochastically crossing the
open detection volume defined by the laser spot lead to a fluctuat-
ing occupation number that follows the Poisson statistics (i.e., the
variance is proportional to the average number of molecules). The
underlying molecular dynamics is extracted as a characteristic decay
time through fluctuation correlation analysis. In its classic view,
FCS is commonly used as alocalmeasurement of the concentration
and characteristic transit time of molecules across the laser beam.
Many efforts targeted the extension of the FCS principle to the
spatial dimension. For instance, the focal area was duplicated [3],
moved in space in laserscanningmicroscopes [4–8], or combined
with fast cameras [9–11]. Using these “spatio-temporal”
approaches, heterogeneity of diffusion constants and concentra-
tions across space was addressed for several molecules on both the
model and actual biological membranes [12, 13]. In the effort to
fill the gap between the FCS and SPT approaches, it was recently
demonstrated that the molecular FCS-based diffusion laws can be
recovered by performing fluctuation analysis at various spatial scales
larger (by spot-variation FCS [14, 15]) or smaller (by STimulated
Emission Depletion, STED [16–19]) than the laser focal area and
then by extrapolating the dynamic behavior of molecules below the
diffraction limit. In all the FCS experiments described, however,
the size of the laser beam is a limit that cannot be overcome. Also,
accurate modeling of the dynamics under study is typically
required. To tackle these issues an alternative method based on
spatio-temporal image correlation spectroscopy (STICS [20]) will
be presented here, which is suitable for the study of the dynamics of
fluorescently tagged molecules on live-cell membranes with high
spatiotemporal resolution (schematic representation in Fig.1). In
particular, TIRF microcopy is exploited to provide accurate optical
sectioning of the plasma membrane, while wide-field imaging by an
EMCCD camera is applied to reach sub-millisecond resolution
(Fig.1a). The spatiotemporal fluctuation analysis proposed here
converts a stack of fluorescence intensity images (Fig.1b) into a
stack of images representing the spatiotemporal evolution of278 Francesco Cardarelli