Fluorescent proteins with known FRET efficiencies used as standards
Gary Boas
How do you know that a gram is a gram or that pH 7 is pH 7? How do
you know that a particular Förster resonance energy transfer (FRET) measurement
is what it claims to be? A group with the National Institute on Alcohol Abuse and
Alcoholism, part of the National Institutes of Health in Bethesda, Md., encountered
this problem when beginning a research project involving FRET. A number of FRET
techniques have been developed in recent decades, but their accuracy is not known.
This, of course, complicates interpretation and comparison of the measurements.
The researchers saw a simple solution: development
of fluorescent proteins with known FRET efficiencies that could be deployed as reference
standards. These could be distributed to other investigators, who could use them
to validate and calibrate their FRET imaging systems. Still, they encountered the
same problem of how to determine the absolute, rather than relative, efficiencies.
Researchers have developed fluorescent protein constructs with known
FRET efficiencies that can be used as reference standards to validate and calibrate
FRET systems. They determined the absolute FRET efficiencies of the proteins by
measuring them with three methods: FLIM-FRET, sensitized acceptor emission FRET
(not shown here) and emission spectra (spectral resonance energy transfer, or sRET).
“It occurred to me that you could
use different methods to measure different aspects of FRET,” said Steven S.
Vogel, principal investigator of the project as well as co-author of a
Biophysical
Journal paper describing the development of the standards, published online
Oct. 13. If each of these yields the same efficiency for a given standard, he noted,
the efficiency is likely accurate.
The group used three methods to measure
the FRET efficiencies of potential standards: one based on fluorescence lifetime
microscopy (FLIM-FRET), one based on sensitized acceptor emission, and another based
on differences in emission spectra.
The standards designed by the researchers
were genetic constructs encoding different fluorescent proteins. Each contained
a single donor and acceptor, separated by progressively larger linkers. They initially
employed cyan and yellow fluorescent proteins. But then “along came a paper
from David Piston’s group” at Vanderbilt University in Nashville, Tenn.,
describing the development of Cerulean, a blue GFP variant that is much brighter
than cyan fluorescent protein. They used this as the donor and evaluated a number
of variants of yellow fluorescent proteins as the acceptor, looking specifically
at quantum efficiency and absorption. Based on these criteria, they determined that
Venus, a yellow-green variant, would be best for the FRET measurements.
In hindsight, they noted an even more
important characteristic of the yellow-green variant: Venus matures and folds very
quickly. A time delay in the acceptor could alter the measured FRET efficiencies,
Vogel explained. “In this case, though, the acceptor would fold faster than
the donor. So as soon as the donor folds, the acceptor is there waiting for it.”
The researchers generated constructs
in which the donor Cerulean was attached to the acceptor Venus with 5, 17 or 32
amino acid linkers and measured FRET efficiencies of HEK 293 cells transfected with
the constructs using the three methods. They performed FLIM-FRET measurements with
a Zeiss laser scanning microscope modified for time-correlated single-photon counting
and fitted with a 40x, 0.8-NA water objective. A Coherent mode-locked Ti:sapphire
laser tuned to 830 nm excited the constructs.
After passing through a series of filters,
emitted photons were detected on a Hamamatsu microchannel plate photomultiplier
and counted and correlated with the excitation laser pulses using a photon-counting
module made by Becker & Hickl GmbH of Berlin. They determined the FLIM-FRET
efficiencies by comparing the average lifetimes of the donors in the presence and
in the absence of the acceptors.
For the sensitized acceptor emission
FRET measurements, they used an Olympus inverted microscope equipped with a 603
oil objective as well as donor, acceptor and FRET filter sets. A 75-W xenon arc
lamp provided excitation, while a 12-bit cooled CCD camera made by QImaging of Burnaby,
British Columbia, Canada, acquired the data.
The researchers performed emission
spectra measurements using a scan head attached to an upright microscope outfitted
with a 20x, 0.5-NA water objective. The scan head and microscope were made by Zeiss.
The Coherent mode-locked Chameleon Ti:sapphire laser, tunable from 710 to 950 nm,
provided excitation. A filter blocked the excitation light, resulting in emissions
in the range of 388 to 719 nm. Spectral images were acquired using all 32 channels
of the microscope’s internal META detector.
The FRET efficiencies were in good
agreement across the three methods, even though two are based on pulsed two-photon
excitation and the other on steady-state excitation. This suggested that they were
accurate. Within each of the methods, the efficiencies for the three constructs
were different, confirming that the researchers could use the constructs as FRET
standards. They subsequently reported FRET standards with efficiencies of 43 ±2,
38 ±3 and 31 ±2 percent.
Vogel hopes that the FRET community
will adopt these standards or even similar ones for validating and calibrating FRET
systems. “There was a first wave of FRET researchers who really understood
the physics as well as the biology. Today, more people are trying to adapt it to
their own individual problems; it’s easy for them to make mistakes without
being aware that they’re doing so. The simple act of making a standard, I
hope, will go a really long way toward getting everybody on the same page.”
The NIH group is continuing the research
that instigated development of the FRET standards: specifically, study of changes
in protein-protein interactions that modulate synaptic efficacy. It also has begun
using another FRET technique — fluorescence anisotropy — and has developed
a variety of control constructs to determine whether it is measuring it correctly.
LATEST NEWS