Angewandte
International Edition
A Journal of the Gesellschaft Deutscher Chemiker
www.angewandte.org
Chemie
Accepted Article
Title:
A Hybrid Orbitrap-Nanoelectromechanical Systems Approach to
Analysis of Individual, Intact Proteins in Real Time
Authors:
Adam P. Neumann, Eric Sage, Dmitri Boll, Maria Reinhardt-
Szyba, Warren Fon, Dmitry Grinfeld, Christophe Masselon,
Sébastien Hentz, John E. Sader, Alexander Makarov, and
Michael L. Roukes
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). The VoR will be published online
in Early View as soon as possible and may be different to this Accepted
Article as a result of editing. Readers should obtain the VoR from the
journal website shown below when it is published to ensure accuracy of
information. The authors are responsible for the content of this Accepted
Article.
To be cited as:
Angew. Chem. Int. Ed.
2024
, e202317064
Link to VoR:
https://doi.org/10.1002/anie.202317064
1
A
H
ybrid Orbitrap
-
N
anoelectromechanical Systems Approach
to
Analysis
of
Individual,
Intact Proteins
in Real Time
Adam P. Neumann
1
,
Eric Sage
1
,
Dmitri Boll
2
,
Maria Reinhardt
-
Szyba
2
,
Warren Fon
1
,
C
h
ristophe
Masselon
3
, S
é
bastien Hentz
4
,
John E. Sader
5
,
Alexander Makarov
2
,
6
,
Michael L. Roukes
1
,
7
*
1
Kavli Nanoscience Institute and Department of Physics, California Institute of Technology,
Pasadena, California 91125, USA
2
Thermo Fisher Scientific,
28199
Bremen, Germany
3
Univ. Grenoble Alpes, CEA, IRIG, Biol
ogie à Grande Echelle, INSERM UA 13, F
-
38054
Grenoble, France
4
Univ. Grenoble Alpes, CEA, Leti, F
-
38000 Grenoble, France
5
Graduate Aerospace Laboratories and Department of Applied Physics, California Institute of
Technology, Pasadena
, California 91125, USA
6
Biomolecular Mass Spectrometry and Proteomics, Bijvoet Center for Biomolecular Research
and Utrecht Institute for Pharmaceutical Sciences, University of Utrecht, Padualaan 8, 3584 CH
Utrecht, The Netherlands
7
Departments of Applied Physics and Bioenginee
ring, California Institute of Technology,
Pasadena, California 91125, USA
* Corresponding Author. E
-
mail: roukes@caltech.edu
10.1002/anie.202317064
Accepted
Manuscript
Angewandte Chemie International Edition
This
article
is protected
by
copyright.
All
rights
reserved.
15213773, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317064 by California Inst of Technology, Wiley Online Library on [31/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2
A
bstract
Nanoelectromechanical systems (NEMS)
-
based mass spectrometry (MS) is an emerging technique
t
hat enables
determin
ation
of
the mass of individual adsorbed particles
,
by driving
nano
mechanical devices at resonance and monitoring
the
real
-
time
changes in their resonance
frequencies
induced by each single molecule adsorption event
.
We incorporate NEMS into a
n
Orbitrap
mass spe
ctrometer
and report o
ur
progress toward
s
leveraging the single
-
molecule
capabilities of the NEMS to enhance
the dynamic range of
conventional
MS instrumentation
and
to
offer new capabilities for
performing deep
proteomic
analysis of clinically relevant sa
mples. We
use the hybrid instrument to deliver
E. coli
GroEL molecules
(801 kDa)
to the
NEMS
devices in their
native, intact
state.
Custom ion optics are used to focus the beam
down
to 40 μm diameter with a
maximum flux of 25 molecules/second.
The mass
spectrum
obtained with NEMS
-
MS
shows
good
agreement with the
known
mass
of GroEL
.
Introduction
Innovations
in
mass spectrometry
(MS)
instrumentation
that
combin
e
high resolution, high
mass accuracy, and high throughput
have
established the position of
MS
-
based
proteomics as the
central
technological
paradigm
driving the post
-
genomic era.
The complete yeast proteome was
quantified
in 2008
,
1
and in 2014 two groups reported the first draft of the human proteome
.
2,3
In
all cases,
these analyses were performed
using Orbitrap instrumentation. The Orbitrap mass
analyzer
,
4
introduced in
2000
,
alleviated the shortcomings of
previous
analyzers, which suffer
ed
from
deleterious factors such as
poor
transmission, low mass accuracy, low speed,
or
limite
d mass
dynamic range
.
Compared with genomic analysis, p
roteomic
s
confronts
a significant increase in
both
sample
complexity and concentration dynamic range
. Humans have
approximately
20,000 genes
2,3
that
encode for millions
of protein forms
.
5
Sample complexity
is further
compounded
by proteolytic
digestion and dissociation
, which is commonly employed
in
“bottom
-
up” analysis to achieve
accurate analyte identification via
tandem
MS spectra.
A further complication is the immense
dynamic range
required for proteomics;
the proteome of
a
HeLa cancer cell
spans
8
orders
-
of
-
magnitude
,
6
while
protein concentrations
i
n the human plasma proteome
span 1
1
orders of
magnitude
.
7
10.1002/anie.202317064
Accepted
Manuscript
Angewandte Chemie International Edition
This
article
is protected
by
copyright.
All
rights
reserved.
15213773, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317064 by California Inst of Technology, Wiley Online Library on [31/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3
The complexity of proteomic samples is often addressed by employing liquid
chromatography
as an orthogonal separation technique at the front
end.
Meanwhile,
there exists
no
protein
analog
to
the
polyme
rase
chain reaction
(PCR ), which is
used
to amplify
DNA and RNA to any quantity.
Accordingly, to achieve deep and complete cover
age of the proteome, analysis with single
-
protein
resolution is essential.
Additionally, at present
MS is
a single
-
channel methodology (providing just
one analysis stream)
,
for which Coulomb repulsion limits the number of molecules that can be
processed si
multaneously.
The result is that
mass spectrometers
, including those using Orbitrap
technology,
fall five to six orders
-
of
-
magnitude short on dynamic range
needed for deep proteome
analysis of clinically relevant samples
.
Current
studies analyzing the deep
proteome are
thus
ensemble averages,
performed on
large samples of cell clones
.
6
Consider
a hypothetical enhancement to MS instrumentation that
increases
the dynamic range
to the theoretical limit
. This would
allow for the entire proteome
from a small sample, such as that
acquired from a single cell,
to be
readily
analyzed within a reasonable timeframe.
This
enhancement
could
take the form of
a
device
or intermediate analyzer
that examines molecules
one
-
by
-
one
.
Here, we report on progress toward
s
the first generation of such
a
n
instrument,
achieved by concatenating
a NEMS array
behind
an Orbitrap mass spectrometer.
Nanoelectromechanical Systems (NEMS) are devices that integrate mechanical functionality
with electrical detection to enable physical measurements
on objects of nanometer dimensions
.
Unlike conventional m
ass spectrometry, NEMS devices enable measurement of mass via changes
in
their
mechanical properties; in particular,
by resolving
the shift in resonance frequency
induced
when
an individual analyte physisorbs upon the device surface
. NEMS are typically con
structed
from materials
such as
monocrystalline silicon or
silicon carbide. Electron beam lithography along
with
nanofabrication processes are
used to form
simple resonant
structures such as
miniature
plates,
cantilevers
,
or beams
, or more complex devices
that include phononic bandgap isolation,
with
typical
feature sizes
of
order
tens of
nanometers to
micrometers. These devices exhibit
attributes
that merit their consideration for MS
:
most
notably,
their extreme
sensitivity to
the
addition of
single
molecules
,
and
their
mass resolution
, which actually increases with increased
mass
within their mass
dynamic
range
that extends well into the hundreds of MDa regime.
8
–
13
This
contrasts mar
kedly with other approaches.
The minute mass of these fabricated structures enables the detection of tiny added particles,
10.1002/anie.202317064
Accepted
Manuscript
Angewandte Chemie International Edition
This
article
is protected
by
copyright.
All
rights
reserved.
15213773, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317064 by California Inst of Technology, Wiley Online Library on [31/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4
such as nanoparticles,
individual proteins
,
protein complexes
, and large biomolecules
in the
hundreds of kDa to
multiple
MDa range,
19,20
and
intact
viruses in the
multiple
MDa range
.
14
–
18
NEMS
also
offer
promise
for
single
-
Dalton resolution
of protein species
.
21
–
23
The princi
ple behind
NEMS
-
MS lies in driving mechanical devices at resonance and monitoring
the
changes
induced
in
their resonance frequencies in real time
as individual analytes physisorb upon their surface
.
19
The
abrupt change in frequency shifts that occur with each adsorption event can t
hus be used to
calculate the mass of each
individual
particle.
The uncertainty in measured mass σ
m
and
the minimum detectable mass of these devices is
proportional only to the frequency fluctuations σ
f
from which the aforementioned frequency shifts
are
extracted. Reducing σ
m
is achieved by operating the NEMS under vacuum and cryogenic
temperatures.
O
ver the measurement
integration
time chosen to
optimally
extract frequency
shifts,
these fluctuations are
observed to remain relatively constant for each de
vice
. Accordingly,
the mass uncertainty of each
analyte measured
is independent of
its
mass. In fact, the value
m
/σ
m
(
referred to as resolution or resolving power
)
actually
increases
with mass, making
NEMS
-
MS
particularly suit
able
for high mass species.
NE
MS
-
MS does not inherently
require ionization of analytes
. E
xperimental realizations of
neutral
-
particle detection with NEMS
19,24
demonstrate they can
circumvent
issues associated with
ion transfer efficiency or heterogeneous peak structure
25
associated with typical MS
measurements
.
NEMS devices are poised to outperform traditional MS in the MDa mass range,
where traditiona
l MS faces challenges
.
25
S
ignificant developments
in
the field of Native MS
,
26
which
permit analysis of
protein complexes in their
intact
state
, have
dramatically extended the mass
range of Orbitrap analysis
,
up to
the scale of
large protein complexes and viruses
,
27
–
29
although
analyte throughput is limited.
Our p
rior inst
antiations of NEMS technology
have
utilized custom
ion optics and vacuum
systems to
direct molecules to the NEMS
, recreating traditional MS
-
based instrumentation
de
novo
.
8,15
These
first instruments
demonstrated
re
latively low
ion
flux
due to the mismatch between
the typical ion beam size of a few
square
mm and the cross
-
section of a typical NEMS resonator at
a few
square
μ
m
.
Work to increase the capture cross section of the ion beam has been
demonstrated with arrays of
up to 20
NEMS devices
,
30
with the possibility of arrays of tens
-
of
-
thousands of devices produced
en masse
utilizing VLSI fabrication techniques.
41
Alternatively, ion
10.1002/anie.202317064
Accepted
Manuscript
Angewandte Chemie International Edition
This
article
is protected
by
copyright.
All
rights
reserved.
15213773, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317064 by California Inst of Technology, Wiley Online Library on [31/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5
optics can be us
ed to significantly shrink the beam size
and to implement scanning of multiple
devices; this permits a single
NEMS
device to capture
a significant fraction of the
analytes
that
approach
it
.
Here, we demonstrate the use of an Orbitrap based mass spectromet
er to deliver
individual,
intact
E. coli
GroEL (801 kDa)
molecules
, a noncovalent protein complex consisting of 14 identical
subunits, to an array of
20
NEMS resonators that
can each
measure the
analyte
mass
one molecule
at a time. Ion optics designed to f
ocus the beam of GroEL ions to a spot size of 50
μ
m
are
manufactured and incorporated into the hybrid instrument
. T
he NEMS array is used to precisely
map the ion flux in three dimensions
. The map thus obtained matches the predicted
focusing
capability and is used to find the optimal location for NEMS operation.
An inertial mass spectrum
of GroEL is compiled from each individual absorption event and the measured mass shows
excellent agreement
with
its expected value.
We discuss future
work
that
will
enable
the
practical
use of NEMS in modern MS instrumentation to extend its capabilities.
Results and Discussion
An overview of the
instrument
is shown in
Figure 1
.
A
Thermo Fisher Scientific
Q
Exactive Plus
instrument
with
Orbitrap
detect
ion
capable of analyzing the mass
-
to
-
charge ratio of species with
extremely high resolving power (up to 140,000 for 200
m/z
)
31
is
m
odified with extended mass
range
(EMR
) capabilities
and
a special quadrupole for selecting
the
species of interest.
The front
end is
outfitted with an electrospray
ionization
(ESI) or nanoelectrospray (n
ano
-
ESI) ionization
source, thus providing a
n
interface for transferring molecular species from solution to the gas
phase
. This
is compatible with
standardized
protocols used
in
proteomi
cs or native mass
spectrometry research.
A similar instrument
was
previously
constructed
and reported
measurement of
intact IgG antibody at 149 kDa, yeast proteasome at 730 kDa, and GroEL at 801
kDa
,
27
and
up to viral nanoparticles at 4.5
MDa
.
28
Suc
cessful realization of the hybrid system require
s
careful
modification
of the design of the
Q
Exactive
MS
and
provision of
the
requisite
operating conditions of both the Q
Exactive and
the
NEMS devices. The Q
Exactive
MS
includes multiple chambers for tran
sferring or trapping ions.
One such chamber, the higher
-
energy collisional dissociation (HCD) cell, includes a removable back
port
that facilitates our
instrument modification
s
.
10.1002/anie.202317064
Accepted
Manuscript
Angewandte Chemie International Edition
This
article
is protected
by
copyright.
All
rights
reserved.
15213773, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317064 by California Inst of Technology, Wiley Online Library on [31/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6
Figure 1 |
Real
-
time detection
of GroEL adsorbed onto
doubly
clamped
NEMS beams in high
vacuum.
a
)
Architecture of the Hybrid Q Exactive
-
NEMS System that delivers intact proteins or
protein complexes to the Orbitrap chamber for analysis of mass
-
to
-
charge ratio and then onto
the NEMS for single molecule analysis.
b
)
SEM image of a 20
-
device array of
doubly clamped
beams shown with silicon, colorized in deep blue, and the metallization layers (Al
:
Si), colored in
yellow, used to interconnect the electrical connections of each resonator
.
c
)
Zoom
-
in on single
in
-
plane r
esonator used in this study with colorized actuation gates and bias electrodes.
d
)
As
GroEL adsorbs onto a NEMS resonator, the resonance frequency of each tracked resonant mode
abruptly shifts, and this shift in resonant frequency is observed in real time.
10.1002/anie.202317064
Accepted
Manuscript
Angewandte Chemie International Edition
This
article
is protected
by
copyright.
All
rights
reserved.
15213773, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317064 by California Inst of Technology, Wiley Online Library on [31/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7
Figure 1A provides a schematic overview of the instrumentation. An additional vacuum chamber
(NEMS Inlet Chamber) is attached to the HCD cell after removing its back plate, followed by a gate
valve and a second vacuum chamber (NEMS Analysis Chamber). The
use of two chambers allows
for differential pumping to maintain the proper vacuum regime for the HCD cell, the Q Exactive MS,
and the NEMS chamber, as well as providing straightforward NEMS device exchange without
breaking the vacuum of the Q Exactive MS.
The HCD cell in the Q
Exactive MS acts as an extended
trapping region for larger molecular weight complexes. Relatively high pressure (
10
−
2
Torr) is
used in this
chamber, both to
trap incoming ions
(
so they can be collected in the C
-
Trap prior to
injection into the Orbitrap
)
and to remove excess water and salt molecules via collisions with gas
molecules
.
27
In contrast
, NEMS operation requires low pressure (
≲
10
−
6
Torr). While dissipation
of NEMS devices due to col
lisions with gas molecules becomes negligible at pressures less than a
few mTor
r
,
32
a larger concern
is
the possib
ility of significant adsorption of ambient gas while
operating the devices at cryogenic temperatures. Such adsorption would lead to a constant drift in
resonance frequency of the devices and enhancement of fluctuation processes that degrade their
Figure 2 |
Ion optics and high
-
resolution map of three
-
dimensional ion flux.
(A)
A series of ion
lenses were designed to focus the ion beam exiting the Q Exactive
MS
into a 50
μ
m spot.
(B)
3D
map of ion flux obtained by extending or retracting the positioner in the Z direction (along the
ion beam) and scanning the NEMS array in X and Y (transverse to the ion beam) such that a data
point is obtained every 20
μ
m. The most focused position in
the second figure has a spot size of
40
μ
m with a peak flux of roughly 25 molecules per second.
a
)
b
)
10.1002/anie.202317064
Accepted
Manuscript
Angewandte Chemie International Edition
This
article
is protected
by
copyright.
All
rights
reserved.
15213773, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317064 by California Inst of Technology, Wiley Online Library on [31/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8
noise performance
.
33,34
A CAD model of the system is provided in
Supporting Information (SI)
Figure 1, and images of the exterior and interior are provided in
SI
Figure 2.
With
this design, the NEMS analysis chamber
is
maintained at a steady pressure of
3
×
10
−
7
Torr
,
regardless of the pressure in the HCD cell.
The end
-
to
-
end gap between transfer quadrupoles
is
set to 5 mm to avoid excessive loss of ion transfer efficiency, and a
miniature gate valve (VAT)
with a gate thickness of 4 mm
is
available to make this possible.
The NEMS chip itself is mounted
on a printed circuit board that rests on a
XYZ
piezoelectric translation
stage
(Attocube)
that
provides
5 mm of travel range
with
sub
-
nanometer positioning precision
along the three axes
. It
is
thermalized to a
flow
cryostat
(Janis)
,
which
is
cooled with liquid nitrogen
.
The system
can
reach
a
base temperature of 80K within 20 hours; cryogenic design and temperature readings are
prov
ided in
SI
Figure 3.
A critical design criterion
is
achieving sufficient ion flux to the NEMS. The typical cross section
for devices used for mass sensing is
of
order of 10
μm
by 0.3
μm
, which represents an extremely
small capture cross section for typica
l ion beam sizes
of
order of a few
square
mm
.
This previously
presented a
major
barrier to obtaining
enough physisorption events to permit adequate
measurement statistics
;
to wit,
a
n
early
ESI based instrument achieved a maximum rate of
only
1
event
per
minute
.
35
.
This limitation
is
surmounted here
with the addition of ion optics immediately
prior to the NEMS
, which greatly
increase
s
the
analyte flux de
nsity delivered to the NEMS
.
Our
design for
a two
-
stage
ion
lens
focusing the ions to a spot size of 50 μm is shown in Figure 2.
Custom
instrument
control
software is used
to
either
trap ions in the HCD cell prior to injecting
them into the O
rbitrap
analyz
er or
to
allow the ions to continue through the HCD cell to the NEMS.
The choice of path can be adjusted rapidly, allowing for Orbitrap
m/z
spectra to be associated with
NEMS measurements and validate
which molecules are being sent to the NEMS. GroEL molec
ules
are ionized using the nano
-
ESI source
(see
SI
)
, and the quadrupole
mass
-
filter
board
was modified
to
filter
molecules with
m/z
greater than 10,000. As shown in
SI
Fig
.
4, a mass spectrum collected
using the Orbitrap analyzer is similar to results reported elsewhere
.
27
The mass spectrum confirms
the GroEL molecules retain their intact 14
-
mer configuration with a calculated mass of
801.
421
kDa
, and the use of the quadrupole filter prevents fragments or other contaminants from reaching
the NEMS.
Our
initial NEMS measurements required significant
effort
to locate the ion beam
,
given
its tight
10.1002/anie.202317064
Accepted
Manuscript
Angewandte Chemie International Edition
This
article
is protected
by
copyright.
All
rights
reserved.
15213773, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317064 by California Inst of Technology, Wiley Online Library on [31/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
9
focusing
implemented in this work
. Prior to col
lecting significant data for MS, we systematically
characterize in three dimensions the ion beam directed toward the NEMS. A 20
-
device NEMS array
is
used for this purpose
, which greatly increases
the throughput of data collection.
A
schematic
diagram of th
e NEMS array operation is provided in
SI
Fig
.
5
, and the electronic operation and
readout scheme is provided in
SI
Fig
.
6
. Additional d
etails concerning the operation and fabrication
of this array are provided in
Ref.
30
. In brief, the devi
ces have a pitch of 20 μm in X and 60 μm in Y
for a total field size of
60
×
240
μm. Devices are 160 nm thick by 300 nm wide; the lengths vary
from 7
–
9.2 μm. Each device possesses five electrical connections. They are interconnected in
parallel using two add
itional metal layers. This ensures that the input signal is applied to all NEMS
simultaneously, and the overall output is the sum of all individual outputs
of the devices
.
Because
the device
resonance frequencies are different, they are separately
frequenc
y
-
addressable.
The
entire array is monitored by sequentially switching between the resonance frequency of each
resonator: a phase
-
locked loop (PLL) locks onto a given resonator using its last known resonance
frequency, registers its current resonance frequ
ency after a given idling time
,
τ (here 10 ms)
,
and
then switches to the next resonator. Frequency time traces for each individual device
are
extracted
after the experiment.
Data
displaying
frequency
-
tracking of myoglobin molecular adsorption
events to eac
h device of the array is shown in
SI
Figure
7
.
To map the ion flux of GroEL in three dimensions, we use the 3D XYZ positioner’s capability for
0.1 nm positioning precision over
a
5 mm travel range. The array has a pitch of 20 μm in X and 60
μm in Y
.
Theref
ore
,
to achieve uniformly spaced data collection, the array
is
first staggered twice
in Y by 20 μm, forming a field of
60
×
300
μm with a data point every 20 μm in X and Y.
This is then
repeated to form a field of
600
×
600
μm perpendicular to the ion beam axis. This
is
done with 3
minutes spent per position, so each
field
takes
1 hour to capture. Finally, it
is
repeated for several
Z values, from fully retracted (further from the lens, Z=125 μm) to fully extended (closer to
the
lens, Z=4995 μm). Positioning
is
performed in an automated manner using a custom Python script.
Results are shown using contour plots in Figure 2 with raw data given in terms of
the NEMS
relative frequency shift
per second.
T
his can be converted to m
olecules
per second
using the mass
responsivity for mode 1
of the NEMS
at 12 Hz/ag
,
or the responsivity for mode 2 (not shown) at
32 Hz/ag
,
30
with
the approximate respective
NEMS
resonance frequencies of 25 and 67
M
Hz, and
a GroEL mass of 1.3 ag. The ions
are
focused to a minimum beam size of 40
μm
, with a maximum
flux rate of 25 molecules
per
s
econd
. Such data, once collected, is useful for finding the ions again
10.1002/anie.202317064
Accepted
Manuscript
Angewandte Chemie International Edition
This
article
is protected
by
copyright.
All
rights
reserved.
15213773, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317064 by California Inst of Technology, Wiley Online Library on [31/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
10
after installing a new
NEMS
device, as well as setting the position of the device to achieve the
desired flux rate.
A similar map
is
obtained with myoglobin ions demonstrating a maximum flux of
880 molecules per second (
SI
Figure
8
).
Following characterization of the ion beam with a
20
-
device array, the smallest device
in the
array
(7 μm length)
wit
h the
lowest (best)
mass resolution
is
operated in isolation at a location
within the focused ion beam
that
provides an
event rate of
~
1 molecule
per second
.
This
rate
is
chosen to demonstrate the ability to collect a large amount of data in a short period
of time, while
still having a low probability of multiple molecules landing within the measurement window of
2
00 ms. This measurement window is set according to the PLL
loop integration
time
(measurement time)
that minimizes the Allan deviation
, which
is
measured
to be
2
.
2
×
10
−
7
for
mode 1 and
1
.
8
×
10
−
7
for mode 2 (with a noise correlation of
휌
=
0
.
28
)
,
corresponding to
an
expected mass resolution of
~
118 kDa.
Jumps are automatically detected and evaluated to build a mass spectrum using
our
previously
reporte
d methods.
8,24
For
a 1D doubly
-
clamped Euler
-
Ber
noulli beam,
the
frequency shifts of the
resonant modes obey:
훿
푓
푛
푓
푛
≈
−
푚
added
2
푀
device
휙
푛
2
(
푥
)
,
(
1
)
where
훿푓
푛
is the change in resonant frequency of the device upon mass adsorption,
푓
푛
is the
resonant frequency of the device without the adsorbate,
푚
added
is the adsorbed mass at position
푥
,
푀
device
is the device mass, and
휙
푛
(
푥
)
is the scaled displacement mode shape of mode
푛
; Eq. (1)
is valid for
푚
added
≪
푀
device
.
As shown in Eq. (
1), the frequency shifts depend not only on mass
but also the (unknown) position where the particle lands. For the doubly
-
clamped beams used in
this study, measurement of the
first two
vibrational modes of the device is sufficient to determine
not only the
unknown position, but also the mass of each particle.
19
The calculation of the mass
spectrum relies on this formula as well as a probability density based formulation through a
change of variables
that
transforms measured frequency shifts and their uncertainties into mass
and posit
ion and their uncertainties.
19,35
Adsorption events
are
identified as fluctuations that
exceed a 5σ deviation from background noise
, which for the first two modes is observed to form
a
bivariate normal distribution over a given measurement window
.
Further d
etails of obtaining the
mass spectrum are provided in
SI
.
The mass spectrum features a primary peak at 7
4
8 kDa, close to
10.1002/anie.202317064
Accepted
Manuscript
Angewandte Chemie International Edition
This
article
is protected
by
copyright.
All
rights
reserved.
15213773, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317064 by California Inst of Technology, Wiley Online Library on [31/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
11
the known mass of 801 kDa
and a
series of
smaller,
unresolved
secondary peak
s at higher mass
(Figure 3)
.
While th
is
mass difference
(~53 kDa)
is close to that of a single GroEL subunit
(~57
kDA)
, previous native mass spectrometry investigations of this molecule, under similar analysis
conditions, did not s
how such fragmentation.
40
Instead, we attribute the mass difference observed
here to
our
uncertainties
(~5%)
in establishing the modal mass
es
of
the
NEMS devices;
accurate
knowledge of
them
is required to deduce analyte mass
precisely
.
A simulation
used to
validate the experimentally obtained spectrum
is performed
as follows.
An 801 kDa point mass is adsorbed with random position to a 1D
doubly clamped
beam.
The
frequency shifts of the resonant modes
are calculated
according to Eq. (1)
for
each
simul
ated
adsorption event,
then
noise
is sampled
from
the experimentally obtained
noise
distribution for
the chosen measurement window and added to the (noise
-
free) modelled frequency shifts.
A mass
spectrum
is
then obtained using 1000 such events according to
the same procedure in
SI
.
The peak
of the simulated spectrum matches the simulation input of 801 kDa. However,
t
he width of the
simulated spectrum, measured by the
full width
at half maximum (FWHM) is 278 kDa
(Figure 3)
,
significantly
smaller
than the exp
erimentally obtained spectrum
, measured
at 445 kDa
.
The relatively broad observed peak compared with the simulation is unlikely to be due to
variation in sample mass
. This is
because the hybrid Orbitrap
-
NEMS instrument allows for
selection of only intact GroEL molecules using the quadrupole mass filter and validation of the
intact nature using the Orbitrap analyzer (
SI
Figure 4).
A doublet
, or two GroEL molecules arriving
at a
single position and
at the same time, should produce a clearly resolvable peak at twice the
main peak, given the device resolution.
This is not observed in the measured spectrum.
One
possibility is a significant fraction of events consist of multiple part
icles landing at different
positions within the chosen measurement interval of 200 ms. Such multiple
-
jump events violate
the assumptions of the probability distribution framework
detailed
in
the
SI
Materials
and would
broaden the final mass spectrum. This
could occur if the true event rate is faster than the estimated
1 molecule per second, or if the molecules do not arrive at a fixed rate but instead in bunches.
Additional work to resolve multiple
-
landing events through use of an improved event
-
detection
s
cheme, or the use of additional modes to detect the presence of multiple particles
36
will
be used
in further studies that
employ
a high event rate.
10.1002/anie.202317064
Accepted
Manuscript
Angewandte Chemie International Edition
This
article
is protected
by
copyright.
All
rights
reserved.
15213773, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317064 by California Inst of Technology, Wiley Online Library on [31/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
12
Conclusion
s
T
h
e
hybrid instrument
we describe
here makes it possible to perform
several important
experiments
.
It
enables
Orbitrap
-
based
verification of the
GroEL ions sent to NEMS devices, and
direct res
olution of
single
-
molecule adsorption events
and subsequent deduction of analyte mass
.
Mass spectra collected with two
simultaneous
modes show
good
agreement with the expected
mass.
I
on optics
is
used to produce a beam with a
diameter
of
40 μm
; this
yields
a maximum
detection rate of 25 molecules per second
for GroEL and 880 molecules per second for myoglobin
.
Using a
20
-
device array
,
consisting of 10 μm by 300 nm
devices
,
currently
permits
detection of 5%
of
the
total
molecul
ar flux
reaching the device ch
ip. This
indicates
that
NEMS can enables a very
low
limit of detection
,
that is,
a low
number of molecules in the initial
solution needed to produce
a signal.
Other
factors affecting the detection limit using
NEMS are
solely
the electrospray
ionization eff
iciency (typically 5
-
10%, with the theoretical limit
up to 85%
39
) and the transfer
efficiency of ions to the NEMS (estimated to b
e 13%
for this hybrid instrument with further details
in
Ref.
40
). Thus,
we estimate that the
current detection limit achieved is
of order
10
4
to
10
5
Figure 3 |
GroEL inertial mass spectrum and simulated spectrum.
a
)
NEMS
-
MS spectrum of
GroEL with
a primary peak at 7
4
8 kDa and
unresolved
secondary peak
s at higher mass
.
b
)
A
simulation matching details of the experiment is shown alongside the observed mass spectrum;
see text for details. In the simulation, the adsorption events are independent, and
the mass
spectrum features a single, narrower peak.
10.1002/anie.202317064
Accepted
Manuscript
Angewandte Chemie International Edition
This
article
is protected
by
copyright.
All
rights
reserved.
15213773, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317064 by California Inst of Technology, Wiley Online Library on [31/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
13
molecules.
T
his enable
s
proteomic experiments on extremely small sample sizes, such as that from
a single cell.
To achieve
significant
improvement to existing
NEMS
-
MS instruments,
two orders
enhancement
of mass
resolution
are needed
,
along with controlled
desorption of individual
molecules from the NEMS devices.
We anticipate that the
former is within reach with
next
-
generation
phononic bandgap NEMS devices with ultralow energy dissipation
.
41,42
These are of
sufficient size to permit analysis of large protein complexes
,
while also potentially
yield
in
g
single
Dalton mass resolution
.
Experiments are underway to explore both soft landing of individual
protein analytes on the NEMS as well as their controlled desorption.
Acknowledgements
The authors gratefully acknowledge support from the Wellcome Leap Fo
undation through its
Delta Tissue program,
the National Science Foundation (Major Research Instrumentation Award
1828787 and Partnerships for Innovation Award 2016555),
and
an
Amgen Chem
-
Bio
-
Engineering
Award (CBEA)
.
We also thank Prof. Albert
Heck
and his
lab
oratory
(Utrecht Univ., NL)
for
helpful
discussions and
provi
sion
of
the GroE
L
samples analyzed.
Author Contributions
M.L.R. and A.A.M. conceived of and directed th
is
effort. A.P.N., E.S., and W.F. built the
nanosystems instrumentation and car
ried out the experiments. D.B., M.R.
-
S., and A.A.M. designed
the specialized ion optics and provided expertise on MS operations. C.M and S.H. provided NEMS
devices and advice on experiments.
A.P.N., J.E.S. and M.L.R analyzed the data.
All authors discussed
the results and edited the paper.
Competing Interests
D.B., M.R.
-
S. and A.A.M. are employees of Thermo Fisher Scientific,
which
commercializes
Orbitrap
-
based
instruments
for MS
.
10.1002/anie.202317064
Accepted
Manuscript
Angewandte Chemie International Edition
This
article
is protected
by
copyright.
All
rights
reserved.
15213773, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317064 by California Inst of Technology, Wiley Online Library on [31/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
14
References
1.
De Godoy, L. M.
et al.
Comprehensive mass
-
spectrometry
-
based proteome quantification of
haploid versus diploid yeast.
Nature
455
, 1251
–
1254 (2008).
2.
Kim, M.
-
S.
et al.
A draft map of the human proteome.
Nature
509
, 575
–
581 (2014).
3.
Wilhelm, M.
et al.
Mass
-
spectrometry
-
based draft of the human proteome.
Nature
509
, 582
–
587 (2014).
4.
Makarov, A. Electrostatic Axially Harmonic Orbital Trapping: A High
-
Performance Technique
of Mass Analysis.
Analytical Chemistry
72
, 1156
–
1162 (2000).
5.
Cho, W. C. Proteomics technologies and challenges.
Genomics, proteomics & bioinformatics
5
,
77
–
85 (2007).
6.
Nagaraj, N.
et al.
Deep proteome and transcriptome mapping of a human cancer cell line.
Mol
Syst Biol
7
, 548
–
548 (2011).
7.
Anderson, N. L. & Ande
rson, N. G. The human plasma proteome: history, character, and
diagnostic prospects.
Molecular & cellular proteomics
1
, 845
–
867 (2002).
8.
Hanay, M. S.
et al.
Single
-
protein nanomechanical mass spectrometry in real time.
Nat Nano
7
, 602
–
608 (2012).
9.
Ekin
ci, K. L., Huang, X. M. H. & Roukes, M. L. Ultrasensitive nanoelectromechanical mass
detection.
Applied Physics Letters
84
, 4469
–
4471 (2004).
10.
B. Ilic
et al.
Attogram detection using nanoelectromechanical oscillators.
Journal of Applied
Physics
95
, 3694
–
3703 (2004).
11.
Yang, Y. T., Callegari, C., Feng, X. L., Ekinci, K. L. & Roukes, M. L. Zeptogram
-
Scale
Nanomechanical Mass Sensing.
Nano Letters
6
, 583
–
586 (2006).
12.
Li, M., Tang, H. X. & Roukes, M. L. Ultra
-
sensitive NEMS
-
based cantilevers for sensing
, scanned
probe and very high
-
frequency applications.
Nat Nanotechnol
2
, 114
–
120 (2007).
10.1002/anie.202317064
Accepted
Manuscript
Angewandte Chemie International Edition
This
article
is protected
by
copyright.
All
rights
reserved.
15213773, ja, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317064 by California Inst of Technology, Wiley Online Library on [31/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License