Focal molography is a new method for the in situ analysis of molecular interactions in biological samples

Abstract

Focal molography is a next-generation biosensor that visualizes specific biomolecular interactions in real time. It transduces affinity modulation on the sensor surface into refractive index modulation caused by target molecules that are bound to a precisely assembled nanopattern of molecular recognition sites, termed the ‘mologram’. The mologram is designed so that laser light is scattered at specifically bound molecules, generating a strong signal in the focus of the mologram via constructive interference, while scattering at nonspecifically bound molecules does not contribute to the effect. We present the realization of molograms on a chip by submicrometre near-field reactive immersion lithography on a light-sensitive monolithic graft copolymer layer. We demonstrate the selective and sensitive detection of biomolecules, which bind to the recognition sites of the mologram in various complex biological samples. This allows the label-free analysis of non-covalent interactions in complex biological samples, without a need for extensive sample preparation, and enables novel time- and cost-saving ways of performing and developing immunoassays for diagnostic tests.

Full text

Focal molography is a new method for the
in situ analysis of molecular interactions in
biological samples
Volker Gatterdam1†, Andreas Frutiger1†, Klaus-Peter Stengele2, Dieter Heindl2,ThomasLübbers
3,
Janos Vörös1*and Christof Fattinger3*
Focal molography is a next-generation biosensor that visualizes specific biomolecular interactions in real time. It
transduces affinity modulation on the sensor surface into refractive index modulation caused by target molecules that are
bound to a precisely assembled nanopattern of molecular recognition sites, termed the ‘mologram’. The mologram is
designed so that laser light is scattered at specifically bound molecules, generating a strong signal in the focus of the
mologram via constructive interference, while scattering at nonspecifically bound molecules does not contribute to the
effect. We present the realization of molograms on a chip by submicrometre near-field reactive immersion lithography on
a light-sensitive monolithic graft copolymer layer. We demonstrate the selective and sensitive detection of biomolecules,
which bind to the recognition sites of the mologram in various complex biological samples. This allows the label-free
analysis of non-covalent interactions in complex biological samples, without a need for extensive sample preparation, and
enables novel time- and cost-saving ways of performing and developing immunoassays for diagnostic tests.
There are two major determinants that have to be fulfilled for
the selective and sensitive detection of biomarkers in
complex biological samples through specific binding (SB) to
recognition sites: (1) high specificity and stability (slow off rate) of
the recognition between the investigated biomolecules and (2) dis-
tinction of the specific recognition from nonspecific binding
(NSB). In heterogeneous assays (for example, the classical enzyme-
linked immunosorbent assay, or ELISA1), the specific molecular
recognition occurs in close proximity to the solid phase, which is
used for immobilization of the recognition element. The solid
phase of a heterogeneous assay serves two fundamental functions:
(1) display of the immobilized recognition site to the sample of
interest and (2) retaining exposure of the recognition site to the
washingsolutionduringthewashingstep.Heterogeneousimmuno-
assays in complex biological samples rely on intensive washing of
the assay surface, which permits the specific binding signal obtained
from the recognition event to be distinguished from signals that
arise from NSB of the assay and sample components. Nanometre-
sized defects in surface order and molecule orientation may
expose unwanted chemical groups at the surface–solution interface,
resulting in undesirable or uncontrollable surface properties
enhancing NSB2. New methods to separate these NSB events from
the molecular recognition are pivotal for selective and sensitive
detection of biomolecules.
In visible light, molecules in aqueous solutions can manifest
themselves either through modulation of the amplitude of the inci-
dent light (absorbing molecules, such as a dye), through modulation
of the phase of the incident light by their refractive index, or by
emitting fluorescent light3. With a few exceptions, biomolecules in
water represent pure phase objects (no absorption) and manifest
themselves only by their refractive index4. Therefore, many optical
biosensors have been developed to measure the amount of proteins
based on refractive index change, which means that all of them also
detect non-specifically bound molecules5–7.
Current state-of-the-art assay methods8typically deal with the
NSB problem by using sandwich assays that allow the discrimi-
nation of the specifically bound secondary binders from the NSB
molecules by extensive rinsing (using the difference in off rate) or
by shear force (which discriminates the different binding strengths)
in so-called force-discrimination assays9,10. Other label-free techniques
like surface plasmon resonance (SPR)11 or Mach–Zehnder12 inter-
ferometry rely heavily on reference channels to obtain meaningful
results. Owing to the spatial separation of the two channels, these
methods require sophisticated set-ups and are also more prone to
variations and gradients. The cantilever-based nanomechanical
detection method of Zhang et al.13 also requires reference probes
and only enables sensitive measurements under static conditions.
The method is not compatible with flow, or detection in a
complex environment like an in situ cell culture measurement.
On the other hand, the detection method described by Stern
et al.14 requires a purification step before detection of the analyte.
It therefore does not allow direct study of the biomolecular
interactions in the native complex sample environment.
There are very few existing concepts that can directly discriminate
between SB and NSB (for example, the biomimetic approach pre-
sented by Cornell et al.15 using ion channels, the molecular diffraction
gratings introduced by Tsay et al.16 and Cleverley et al.17,orthe
recently introduced single-molecule techniques such as kinetic finger-
printing18 that discriminate SB from NSB based on residence time and
corresponding statistical parameters19). However, none of these are
available in the current diagnostic market, possibly owing to the
complexity of the required surface architecture or instrumentation.
Here, we present a realization of focal molography, a technique
that can discriminate SB and NSB and is also sensitive and simple
1Laboratory of Biosensors and Bioelectronics, Institute for Biomedical Engineering, ETH Zurich, 8092 Zurich, Switzerland. 2Roche Diagnostics GmbH, 82377
Penzberg, Germany. 3Roche Pharma Research and Early Development, Roche Innovation Center Basel, 4070 Basel, Switzerland. †These authors contributed
equally to this work. *e-mail: janos.voros@biomed.ee.ethz.ch;christof.fattinger@roche.com
ARTICLES
PUBLISHED ONLINE: 25 SEPTEMBER 2017 | DOI: 10.1038/NNANO.2017.168
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology 1
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

enough to potentially make it to the diagnostic market. The concept
of focal molography was described recently20. Briefly, in focal molo-
graphy, coherent laser light is diffracted at a coherent assembly of
interaction sites (Fig. 1a). The coherent assembly comprises a
precise submicrometre pattern of identical interaction sites that
needs to be phase-matched to the incident coherent light over the
entire pattern. The light diffracted at this assembly constructively
interferes in a diffraction-limited focal spot (Airy disc).
A molecular assembly with these properties is termed a
‘mologram’. In essence, a mologram is a synthetic hologram
composed of molecules, which scatters light into a diffraction-
limited focal spot and is tuned such that it allows molecular inter-
action analysis. Acquisition of the molographic signal is realized
through detection of the diffracted coherent light in the focus of
ONH
O
HN
ONH
O
HN
O
O
O
NH2
NH2
NO2
OHHO
PEG2,000
PEG3,400
n
ONH
O
HN
ONH
O
HN
O
O
O
NH
NH2
NO2
OHHO
PEG2,000
O
O2N
O
S
PEG3,400
n
O
NH
O
NH
O
O
O
N
NH
PEG
2,000
O
O
O
O
O
O
O
O
O
2
2
2
2
2
N
N
N
N
N
O
O
O
O
O
S
S
S
PEG
3,40
0
ONH
O
HN
ONH
O
HN
O
O
O
NH
NH2
NO2
OHHO
O
S
NH
HN
O
PEG2,000
PEG3,400
n
Laser beam
Guided mode
Photo detector
array
Sensor chip
Mologram
Waveguide
a
400 μm
c
(iv)
(iii)
(ii)
(i)
bPhSNPPOC
390 nm
Full field illumination
Biotinylation
Grooves ‘inactive‘
Ridges ‘active‘
SAv555
210–148 nm
Graft copolymer
Ta 2O5
390 nm
420–296 nm
(i)
(ii)
(iii)
(iv)
(v)
(vi)
Phase mask
Liquid
Glass substrate
Glass substrate
Figure 1 | Schematic illustration of the fundamental components of focal molography and the submicrometre near-field lithographic process used in
molography. a, The sensor chip is based on a single-mode optical waveguide with a grating coupler. The guided mode is diffracted at the mologram and
focused to a diffraction-limited spot. The central curved recess area in the mologram prevents the second-order Bragg reflection (for details see ref. 20).
The spot intensity, which correlates quadratically with the bound mass of the analyte, is detected by a photodetector array. For details of the chip and
readout configuration see Supplementary Figs 1 and 2. b, For a better understanding, the copolymer structures at different stages of the process are
connected via dashed lines. (i) The 145 nm metal oxide (Ta
2
O
5
) layer is coated with a thin photosensitive graft copolymer layer containing PhSNPPOC
protected amino-PEG
3400
polymer. (ii) Phase mask technology is used for activation of the monolithic photosensitive layer. The physics of the phase mask
lithography process results in structures with half the period of the phase mask (420–296 nm). This allows the creation of the nanostructures (210–148 nm
line width) of the mologram: [NH
2
|NH-PhSNPPOC] (see Supplementary Section ‘Nomenclature of affinity modulation and surface composition of
molograms’). (iii) Induced activation contrast after photolithography. Activated areas are termed ‘ridges’and inactive areas ‘grooves’.(iv)Further
functionalization with amine reactive compounds leads to the desired binding properties (chemical functionalization). NHS-biotin for streptavidin (SAv)
binding: [NH-biotin|NH-PhSNPPOC]. (v) To minimize the difference in NSB between grooves and ridges, or to realize backfilling of the mologram, the
remaining PhSNPPOC groups are photocleaved. Additional passivation (not shown) can be obtained by amine reactive blocking reagents, for example,
NHS-PEG. (vi) The molographic signal is generated by binding of SAv: [NH-biotin/SAv|NH
2
]. c, Images showing the molographic raw data at four exemplary
time points. The drawn molograms are an illustrated overlay and are invisible in reality. The middle row of the mologram overlay is 90° rotated and does not
generate a signal. Because this pattern is incoherent with regard to the direction of the guided wave, the structure generates no diffractive signal, although,
biochemically, the same molecular interaction takes place.
ARTICLES NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2017.168
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology2
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

the mologram. The light intensity in the molographic focus scales
quadratically with the surface mass density modulation Δ
Γ
of the
mologram. Background scattering by molecules that are randomly
attached to the assay surface is intrinsically separated by focal
molography from the coherent molographic signal. In addition,
molography does not require the presence of a label, as in ELISA
assays. Label-free molography visualizes biomolecular recognition
events through the detection of minute phase-contrast changes in
the nanoenvironment of the recognition sites in the mologram20.
This opens up new perspectives and possibilities in analysing non-
covalent interactions and especially in the discovery of unknown bio-
molecules in complex biological samples by pulldown assays20.
In the challenging case of measuring biological responses and
interactions directly in living tissue or cell cultures, most bioanalytical
methods face severe problems with background signals or involve
laborious referencing procedures. Although some methods (fluor-
escence-based techniques such as Förster resonance energy transfer
(FRET) or the radiolabel-containing LigandTracer technology21,22)
bypass some of these problems, they may disturb the studied molecu-
lar interactions as a result of staining or other interfering labels. In
contrast, focal molography transforms the invisible affinity modu-
lation of the mologram into refractive index modulation, inducing a
focused beam of light20,23, where the affinity modulation Δ
A
is the
difference in recognition site density for the target molecule
between ridges and grooves. The sensitivity of most current biosen-
sors is fundamentally limited by NSB events24, but this novel analyti-
cal method enables the collective and sensitive discrimination of
specific biomolecular recognition from NSB on a solid surface in het-
erogeneous assays. Focal molography achieves this by minimizing
separation of the sensing and reference channel to its extreme—to
less than the wavelength of light—while several hundreds of referen-
cing and signal areas are embedded into a single sensing spot.
Here, we present the experimental realization of focal molo-
graphy and the synthesis of molograms on a biosensor chip by
submicrometre near-field reactive immersion lithography (RIL) on
a light-sensitive monolithic graft copolymer layer. For the first
time, we demonstrate an emerging molographic signal that evolves
in real time from a weak background (Fig. 1c and Supplementary
Movie). We have developed a surface chemistry that enables the
fabrication of molograms on an optical waveguide by using a
simple-to-use phase mask process, which allows submicrometre
photopatterning of a non-fouling graft copolymer under immersion
on a laboratory bench. This allows the tuning of surface chemistries
specific to the desired analytical application. We also characterized
the lithography process in detail with simulations and stimulated
emission and depletion (STED) microscopy.
To demonstrate the capabilities of focal molography we will
outline pivotal experiments with complex biological samples. We
will first show that the technology has all the capabilities of refracto-
metric sensors by performing a classical sandwich immunoassay.
Second, we will describe a label-free assay (β-amyloid peptide)
where we compare the binding behaviour in buffer and diluted
serum. Third, we will demonstrate the unique properties and clinical
relevance of focal molography by acquiring real-time label-free
binding curves of a therapeutic antibody in human plasma
samples. Finally, we have recorded the production of antibodies in
a hybridoma cell culture in situ over long time periods. These are
complex experiments that are hard to achieve using other label-
free detection methods, which mainly can only measure large
changes (that is, cell growth or morphology). In addition, state-of-
the-art label-free detection methods often fail to successfully
perform biomolecular interaction analysis in the presence of large
amounts of background proteins or living cells. However, thanks
to the inherent self-referencing of molography, we were able to
obtain meaningful label-free data of molecular binding over long
time periods in complex solutions, even without temperature
stabilization25. For these reasons, we believe that focal molography
has the potential to become the future label-free method of choice
to study a large variety of biomolecular recognition and interaction
processes in complex biological media on a sensorchip in real time—
for example, the relation between protein secretion and cellular
responses such as apoptosis26,27.
(i) (ii) (iii)
c
a
Time (min)
0 5 10 15 20 25
Time (min)
×104
(ΔΓ (pg mm–2))2
(i) (ii) (iii)
0 5 10 15 20 25
0
5
10
15
20
25
30
35
Surface mass (pg mm–2)
OWLS signal
Molographic signal
ΔΓ (pg mm–2)
(i) (ii) (iii)
0
1,000
2,000
3,000
4,000
0
200
400
600
b
Figure 2 | Comparison of sandwich immunoassay performed by molography and OWLS (diffractometric versus refractometric biosensor).
a, (i) Immobilization of SAv and biotinylated anti-TSH F(ab′)
2
. (ii) Immobilization of the target hormone TSH. (iii) Signal enhancement by a second anti-TSH
IgG. (For a full explanation of data processing, see Supplementary Figs 4 and 5). The signal consists of nine molograms represented with the barely visible
standard deviation (s.d.) as a grey ribbon. b, The linear signal overlaps with the OWLS signal at multiple key points. c, Schematic representation of the
performed immuno-sandwich assay.
NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2017.168 ARTICLES
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology 3
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

To perform a meaningful molographic measurement, a few
instrumentation requirements need to be satisfied. Focal molography
requires a suitable dark-field illumination of the mologram, which
was realized here by the guided mode of a high-refractive-index
thin-film optical waveguide (Fig. 1a)20,23. This assures a high
signal-to-background ratio. Molography also needs a submicro-
metre-scale coherent pattern of recognition sites, which can be
created by a variety of existing biological lithography methods that
were developed for biochips28. Here, this coherent pattern was fab-
ricated by photolithography, as described for biotin29,30, nanoparti-
cles31 or peptide synthesis32. For molography we used a new
photoactivatable graft copolymer, similar to the graft copolymers
recently described by Serrano et al.33, to coat the tantalum pentoxide
(Ta
2
O
5
) thin-film optical waveguide34 (see Supplementary
Information). The novel copolymer that we designed for mologra-
phy has four key properties. (1) Before photolithographic exposure,
the copolymer layer on the assay surface is monolithic and homo-
geneous, both in terms of its chemical and optical properties. The
monolithic layer does not contain coherently arranged molecules,
so it shows no molographic signal before the photolithography
step. (2) The copolymer layer is resistant to nonspecific protein
adsorption. However, as with any other protein-repellent surface,
this copolymer layer exhibits defects and irregularities on the nano-
metre scale, which cause residual NSB to the assay surface2,35.
Because these defects are not coherently arranged, they do not con-
tribute to the molographic signal. (3) The copolymer layer contains
a photoactivatable functional group for formation of the recognition
sites by exposure to light. It is important to note that photolithogra-
phy at a wavelength of 390 nm, at the energy doses used, does not
induce coherent defects in the copolymer layer. (If it did, we
would measure a signal from NSB to photo-induced defects in the
whole-blood experiment presented in Supplementary Fig. 10a.)
(4) The formed copolymer adlayer allows necessary chemical and
biochemical modifications after initial surface coating.
The molographic pattern on the assay surface was created by
submicrometre near-field RIL at 390 nm wavelength36 (Fig. 1b
and Supplementary Fig. 1). This wavelength is compatible with
biological systems; for example, it does not induce structural
changes in biomolecules. The submicrometre resolution of the
photolithography step was realized by means of phase-mask
technology37,38. Illumination through the phase mask results in
activated curved lines called ‘ridges’, and inactive lines, in
between, termed ‘grooves’. Using this method, a coherent
molographic assembly of recognition sites can be generated on the
assay surface. For a more detailed characterization of the mologram
and an explanation of the mologram’s nomenclature, see
Supplementary Section ‘Nomenclature of affinity modulation and
surface composition of molograms’.
Sandwich immunoassay
As a proof of principle that the technology has all the capabilities of
state-of-the-art label-free refractometric sensors, we detected
thyroid stimulating hormone (TSH) in a label-free, as well as a
sandwich-type, manner. For comparison purposes, the experiment
was conducted in parallel on a refractometric optical waveguide
lightmode spectroscopy (OWLS) sensor system39. To allow for
real-time observation and determination of the binding kinetics, a
fluidic cell was designed for the following experiments
(Supplementary Fig. 3b). Figure 2 shows the performed immuno-
assay based on molographic (Fig. 2a) and OWLS detection
(Fig. 2b) as well as a schematic of the assay steps (Fig. 2c). A
SAv/biotinylated F(ab)
2
precomplex was first bound to a
0204060
c(IgG) (nM)
0 150 300 4500 150 300 450
0
25
50
75
100
Δ
Γ
(pg mm–2)
Time (s) Time (s)
66.6
33.3
16.6
6.6
3.3
0.67
0
IgG (nM)
Sample Buer
0
25
50
75
100
Δ
Γ
(pg mm–2)
Δ
Γ
(pg mm–2)
Sample Buer
ab
0
25
50
75
100
KD
c
Figure 3 | Determination of assay specificaffinity for antibody binding to β-amyloid peptide. a, Injection of an antibody against immobilized β-amyloid
peptide, followed by washing with buffer. b, Repetition of the experiment with antibody in 20% horse serum. Both data sets show the raw unreferenced
molographic signal. c, Determination of K
D
in running buffer and serum. Endpoint values of the binding experiment in PBS-T buffer (t= 450 s) are represented
by black bars and in serum by red bars in PBS-T buffer. Error bars represent the standard error of the mean (s.e.m.) of eight molograms. c, concentration.
ab
−10 0 10 20 30 40
0
50
100
150
200
400
100
25
6.25
1.563
0
Time (min)
Δ
Γ
(pg mm–2)
Δ
Γ
(pg mm–2)
IgG (nM)
Sample Buer
c(IgG) (nM)
#1
#2
#3
0
50
100
150
200
0.1 1 10 100 1,0000
LOD
Figure 4 | Real-time detection of a therapeutic antibody in human plasma. a, Association and dissociation curves of a therapeutic antibody against an
immobilized idiotypic antibody, displayed as mean signals of at least eight molograms normalized to the bound signal of the idiotypic antibody. The data are
not baseline-corrected and therefore represent the raw data signal. b, Logarithmic representation of therapeutic antibody concentration against endpoint
values for determination of LOD at three times the s.d. of the endpoint value of three different donor blanks.
ARTICLES NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2017.168
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology4
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

[NH-biotin|NH-PEG
12
] mologram, followed by immobilization of the
target hormone TSH and the secondary detection antibody. On taking
thesquarerootofthemolographicsignal,thetwocurvesvirtually
overlap (Fig. 2b), illustrating the expected quadratic dependency of
the molographic signal with the surface mass density modulation
and hence the amount of bound molecules20.Forpracticalreasons,
we refer to Δ
Γ
as the mass density modulation of the mologram. For
an ideal canonic mologram20, the value of Δ
Γ
would be equal in
size to the surface mass determined by OWLS. In reality, the modu-
lation is smaller and therefore also the correlating Δ
Γ
value. The
OWLS signal after SAv binding (not shown) was used for normaliza-
tion of the STED fluorescence signal, and the surface mass density
modulation (283 pg mm
–2
from STED) was used to normalize the
molographic signal. A signal stability test and a negative control
of the TSH assay (lacking the primary capture antibody) are
presented in Supplementary Fig. 11. The key feature of robustness
against incoherent influences such as temperature, buffer composition
(Supplementary Fig. 11) and NSB makes molography promising for
diagnostic applications.
Determining specificaffinity in a direct binding assay in serum
Analysis of biomolecular interactions is one of the most important
areas in life science. This is why label-free biosensors such as SPR
are widely used in biochemistry and biology laboratories.
However, routinely, these methods only work with purified
samples. Samples that are more complex, like blood plasma,
require a high amount of optimization and working time. Focal
molography can offer a better option to obtain unique data on
molecular interactions in biologically relevant solutions. As an
example, we present a kinetic analysis between an antibody and the
β-amyloid peptide, which is clinically relevant for the treatment of
Alzheimer’s disease with active or passive vaccination40. We demon-
strate that the dissociation constant K
D
of an antibody against β-
amyloid can not only be determined in buffer, but also in serum.
We injected different concentrations of the antibody in a flow cell
containing a β-amyloid peptide modified assay surface (Fig. 3a).
The expected concentration-dependent increase in the molographic
signal intensity was observed. After 5 min, the association phase
was interrupted, and buffer was introduced into the flow cell. The
expected slow dissociation for the higher antibody concentrations
was observed, while a small signal increase was found for lower con-
centrations. This can be explained by the non-optimized geometry of
the used flow cell, because residual antibodies from corner areas were
only slowly washed out of the system by the buffer. An identical
experiment with the antibody spiked into 20% horse serum revealed
a comparable binding behaviour (Fig. 3b). A K
D
of 10.1 ± 0.5 nM in
PBS-T and 14.1 ± 1.8 nM in serum could be derived from the
equilibrium binding state (Fig. 3c). This first molographic example
of real-time target binding and affinity determination in a complex
biological fluid was performed without any referencing or additional
surface-blocking steps. It therefore gives only a glimpse of what can be
achieved in the future by focal molography.
Detection of a therapeutic antibody in human plasma
To show the usefulness of the technology in a potential clinical
application we detected, in a label-free manner, the concentration
0 5 10 15 20
0
250
500
750
1,000
1,250
Time (h)
Δ
Γ
(pg mm–2)
02468
0
250
500
750
1,000
1,250 (i) (ii) (iii) (iv)
Δ
Γ
(pg mm–2)
Time (h)
Suspension cell culture
Sensor chip
Molograms
ac
bd
0
25
50
75
100
125
Cell
medium
0.5 × 106
cells per ml
ns
****
****
[SAv/ProteinA|PEG]
[SAv/|PEG]
(×104)
2.0 × 106
cells per ml
(
Δ
Γ
(pg mm–2))2
Figure 5 | In situ measurement of secreted immunoglobulins in a hybridoma cell culture. Four identical molographic chips were functionalized with
biotinylated Protein A and non-functionalized (SAv) in separate wells. a, Real-time mass density increase during in situ measurement. Comparing the signal
increase between the chip with (black curve, 2.0 × 10
6
cells ml
−1
; grey curve, 0.5×10
6
cells ml
−1
) and without (dashed curve) cells shows the production of
proteins that bind to Protein A. The random signal spikes are artefacts due to imprecise repositioning of the instrument and hence different in-coupling
efficiencies of the grating coupler while serial scanning of the different flow channels and chips. b, Zoom-in to the first part of the real-time binding curve,
which was used for SAv normalization of the different chips (i). In half of the flow chambers, SAv was removed by DMSO injection and substituted with
SAv/Protein A (ii). At time point (iii) the cell culture medium was added and at (iv) the cells were added. c, A bright-field microscopy image demonstrating
the compatibility of molography with a high cell density on the sensor surface. Scale bar, 40 μm. d, Mass density increase by endpoint value measurement.
Comparison of multiple endpoint values of a Protein A and a non-functionalized SAv molographic surface. The statistical difference for the cell-containing
chips was significant. ****P <0.0001; NS, not significant.
NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2017.168 ARTICLES
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology 5
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

of a therapeutic antibody in a diluted human plasma sample.
Assessing the therapeutic antibody concentration is key to deter-
mining the physiological half-life of a biopharmaceutical in the
bloodstream of a patient. In our pioneering experiment, we fabri-
cated a molographic chip containing an idiotypic capture antibody
recognizing a therapeutic antibody. Human plasma samples spiked
with different concentrations of this therapeutic antibody were used
as a model system for potential patient samples. The compatibility
of the molographic system with higher concentrations of the
initial complex samples, in this case 10% plasma, enables a higher
sensitivity towards the overall detection of low-abundant plasma
proteins. Injection of diluted plasma samples exhibited the expected
concentration-dependent increase in the molographic signal inten-
sity (Fig. 4). Evaluation of the reproducibilityof the measurements is
summarized in Supplementary Table 3. It is noteworthy that the
presented data for the β-amyloid and the therapeutic antibody are
not reference-corrected, unlike the common practice for established
biosensing systems41. Because, for refractometric biosensors, the
signals caused by NSB are substantially larger than the specific
binding signal, subtraction of NSB from the detected signal
usually completely obscures the specific binding signal for clinically
relevant concentrations of the analyte. In addition, the variation of
NSB from patient to patient does not allow sensors to be calibrated,
rendering most of the established techniques non-applicable for
diagnostic tests42. Therefore, the intrinsic self-referencing, in combi-
nation with the low NSB of the copolymer surface, gives focal molo-
graphy a unique advantage for operation in complex samples.
Thanks to this robustness, and the possibilities for multiplexing,
miniaturization and wash-free operation, one could even envision
performing such interaction analyses with individualized treatment
for every patient. For the underlying assay, the limit of detection
(LOD) was determined at the stable binding phase during
washing of the formed immune complex (Fig. 4a). Evaluation of
the maximal binding signal in correlation with the concentration
resulted in a reproducible LOD of 200 ng ml
–1
(1.3 nM) as the
detectable antibody concentration in human plasma (Fig. 4b).
This is already similar to the performance of the best published
data with state-of-the-art label-free SPR instruments for antibodies
and cancer markers, which mainly determine the LOD in buffer25,43.
The LOD is a relevant parameter for every diagnostic assay, as it
defines the lowest analyte concentration that can be distinguished
from the absence of the substance within a confidence limit
(usually 99% or three sigmas)44,45. For a distinct immunometric
assay, the LOD is expressed in terms of sample concentration
(that is, ng ml
–1
), whereas an assay independent LOD that charac-
terizes the capabilities of the sensor technology can be obtained as
the surface mass density modulation between ridges and grooves on
the mologram (that is, pg mm
–2
)46,47. A surface load of 1 pg mm
–2
corresponds to ∼1 RU (RU: response unit) on the Biocore system
(the gold standard for label-free detection; see ref. 20 and references
therein). Scientists dealing with a broad variety of label-free sensor
devices often use ‘refractive index units’(RIU) instead of RUs. A
sensor resolution of 10
−6
RIU means that a refractive index
change of 10
−6
in the liquid on the sensor surface can be measured
reliably. A surface load of 1 pg mm
–2
( 1 RU) in the aqueous
running buffer corresponds to a change in RIU of ∼2×10
−6
RIU
(see page 3 of ref. 20). The assay independent LOD of the molo-
graphic system was determined as three timesthe standard deviation
of the molographic baseline before analyte injection. The presented
experiment with three individual chips achieved a sensitivity of
∼5pgmm
–2
. This value was obtained through normalization with
the STED data and therefore only provided an upper limit bound
of the LOD. These LOD values were already reasonable (the
detection limit of SPR is currently between 0.1 and 1 pg mm
–2
and was at 10–20 pg mm
–2
in the early 1990s when the label-free
SPR method was introduced)25,47. In such complex samples in
particular, we expect that the LOD of focal molography can be
improved by at least two orders of magnitude by further optimizing
the lithography process and by using an optimized set-up for the
molographic readout.
In situ cell culture measurements
Although several (some commercial) tools exist for monitoring the
behaviour of living cells in an ensemble, these typically only provide
information about morphological changes, and lack molecular
specificity48. However, the robustness of focal molography allowed
long-lasting, in situ measurements of secreted antibodies (IgG)
produced by a hybridoma cell culture. Protein A, a 42 kDa surface
protein originally found in the cell wall of the bacterium
Staphylococcus aureus, was used as the binding partner to
measure the secreted antibody, because it contains four high-affinity
binding sites capable of interacting with the Fc region from IgG. The
cells were cultured on the molographic chip in an open version of
the flow chamber containing 500 µl of the cell culture (Fig. 5c).
The sensor chip consisted of two separate mologram sets:
[NH-biotin/SAv/Bi-Protein A|NH-PEG
12
] (antibody-capturing molo-
gram with biotinylated Protein A (Bi-Protein A)) and [NH-biotin/
SAv|NH-PEG
12
] (negative control). The chip was placed inside the
ZeptoReader and the molographic signal was measured on a
regular basis over 24 h. The experiment was performed with differ-
ent cell densities and in replicates of at least three. As expected,
different molographic responses could be seen for the negative
control compared to the different cell densities (Fig. 5a,b,d).
Conclusion
In summary, we have demonstrated the realization of focal molo-
graphy as a new bioanalytical technique, which distinguishes itself
from other techniques by its insensitivity to NSB, resulting in
robustness and wide applicability. In all studies that investigate
molecular recognition and binding, the molographic biosensor
principle is of great interest because it allows for label-free detection
of trace amounts of all kinds of biomolecules (for example,
hormones, chemokines, cytokines and interleukins) in real
biological samples (for example, serum, plasma or the supernatant
of in vitro cell culture experiments). This new method thus opens
up possibilities in many areas of biology, such as the monitoring
of cell–cell communication, secretion of biomolecules and also
cell surface interactions. Owing to its simplicity, robustness and
potential for miniaturization, focal molography will also have an
impact on the development of future hand-held point-of-care
diagnostic devices.
Methods
Methods and any associated references are available in the online
version of the paper.
Received 13 January 2017; accepted 11 July 2017;
published online 25 September 2017; corrected online
25 October 2017
References
1. Engvall, E. & Perlmann, P. Enzyme-linked immunosorbent assay (ELISA)
quantitative assay of immunoglobulin G. Immunochemistry 8, 871–874 (1971).
2. Love, J. C., Estroff, L. A., Kriebel, J. K., Nuzzo, R. G. & Whitesides, G. M.
Self-assembled monolayers of thiolates on metals as a form of nanotechnology.
Chem. Rev. 105, 1103–1169 (2005).
3. Shimomura, O. The discovery of aequorin and green fluorescent protein.
J. Microsc. 217, 3–15 (2005).
4. McMeekin, T. L., Groves, M. L. & Hipp, N. J. in Amino Acids and Serum Proteins
Vol. 44 (ed. Stekol, J. A.) 54–66 (American Chemical Society, 1964).
5. Vörös, J. The density and refractive index of adsorbing protein layers. Biophys. J.
87, 553–561 (2004).
6. Homola, J. Present and future of surface plasmon resonance biosensors.
Anal. Bioanal. Chem. 377, 528–539 (2003).
ARTICLES NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2017.168
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology6
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

7. Brolo, A. G. Plasmonics for future biosensors. Nat. Photon. 6, 709–713 (2012).
8. Wilson, R. Sensitivity and specificity: twin goals of proteomics assays. Can they
be combined? Expert Rev. Proteomics 10, 135–149 (2013).
9. Lee, G. U., Metzger, S., Natesan, M., Yanavich, C. & Dufrene, Y. F.
Implementation of force differentiation in the immunoassay. Anal. Biochem.
287, 261–271 (2000).
10. Mulvaney, S. P. et al. Rapid, femtomolar bioassays in complex matrices combining
microfluidics and magnetoelectronics. Biosens. Bioelectron. 23, 191–200 (2007).
11. Myszyka, D. G.Improving biosensor analysis. J. Mol. Recognit. 12, 279–284 (1999).
12. Kozma, P., Kehl, F., Ehrentreich-Förster, E., Stamm, C. & Bier, F. F. Integrated
planar optical waveguide interferometer biosensors: a comparative review.
Biosens. Bioelectron. 58, 287–307 (2014).
13. Zhang, J. et al. Rapid and label-free nanomechanical detection of biomarker
transcripts in human RNA. Nat. Nanotech. 1, 214–220 (2006).
14. Stern, E. et al. Label-free biomarker detection from whole blood. Nat. Nanotech.
5, 138–142 (2010).
15. Cornell, B. A. et al. A biosensor that uses ion-channel switches. Nature 387,
580–583 (1997).
16. Tsay, Y. G. et al. Optical biosensor assay (OBA). Clin. Chem. 37,
1502–1505 (1991).
17. Cleverley, S., Chen, I. & Houle, J. F. Label-free and amplified quantitation of
proteins in complex mixtures using diffractive optics technology. J. Chromatogr.
878, 264–270 (2010).
18. Johnson-Buck, A. et al. Kinetic fingerprinting to identify and count single
nucleic acids. Nat. Biotechnol. 33, 730–732 (2015).
19. Gunnarsson, A. et al. Drug discovery at the single molecule level: inhibition-in-
solution assay of membrane-reconstituted β-secretase using single-molecule
imaging. Anal. Chem. 87, 4100–4103 (2015).
20. Fattinger, C. Focal molography: coherent microscopic detection of biomolecular
interaction. Phys. Rev. X 4, 031024 (2014).
21. Gedda, L., Björkelund, H. & Andersson, K. Real-time immunohistochemistry
analysis of embedded tissue. Appl. Radiat. Isot. 68, 2372–2376 (2010).
22. Bondza, S., Stenberg, J., Nestor, M., Andersson, K. & Bjorkelund, H. Conjugation
effects on antibody–drug conjugates: evaluation of interaction kinetics in real
time on living cells. Mol. Pharm. 11, 4154–4163 (2014).
23. Binnig, G. Coherent signal picks out biomolecular interactions. Physics 7,
84 (2014).
24. Woolley, C. F., Hayes, M. A., Mahanti, P., Douglass Gilman, S. & Taylor, T.
Theoretical limitations of quantification for noncompetitive sandwich
immunoassays. Anal. Bioanal. Chem. 407, 8605–8615 (2015).
25. Homola, J. Surface plasmon resonance sensors for detection of chemical and
biological species. Chem. Rev. 108, 462–493 (2008).
26. Ashkenazi, A. & Dixit, V. M. Death receptors: signaling and modulation. Science
281, 1305–1308 (1998).
27. Jemmerson, R., LaPlante, B. & Treeful, A. Release of intact, monomeric
cytochrome cfrom apoptotic and necrotic cells. Cell Death Differ. 9,
538–548 (2002).
28. Jonkheijm, P., Weinrich, D., Schröder, H., Niemeyer, C. M. & Waldmann, H.
Chemical strategies for generating protein biochips. Angew. Chem. Int. Ed. 47,
9618–9647 (2008).
29. Pirrung, M. C. & Huang, C. Y. A general method for the spatially defined
immobilization of biomolecules on glass surfaces using ‘caged’biotin. Bioconjug.
Chem. 7, 317–321 (1996).
30. Sundberg, S. A. et al. Spatially-addressable immobilization of macromolecules
on solid supports. J. Am. Chem. Soc. 117, 12050–12057 (1995).
31. Vossmeyer, T., DeIonno, E. & Heath, J. R. Light-directed assembly of
nanoparticles. Angew. Chem. Int. Ed. 36, 1080–1083 (1997).
32. Fodor, S. P. et al. Light-directed, spatially addressable parallel chemical synthesis.
Science 251, 767–773 (1991).
33. Serrano, Â., Zürcher, S., Tosatti, S. & Spencer, N. D. Imparting nonfouling
properties to chemically distinct surfaces with a single adsorbing polymer: a
multimodal binding approach. Macromol. Rapid Commun. 37, 622–629 (2016).
34. Schmitt, K., Oehse, K., Sulz, G. & Hoffmann, C. Evanescent field sensors based
on tantalum pentoxide waveguides—a review. Sensors 8, 711–738 (2008).
35. Pasche, S., De Paul, S. M., Vörös, J., Spencer, N. D. & Textor, M. Poly(L-lysine)-
graft-poly(ethylene glycol) assembled monolayers on niobium oxide surfaces: a
quantitative study of the influence of polymer interfacial architecture on
resistance to protein adsorption by ToF-SIMS and in situ OWLS. Langmuir 19,
9216–9225 (2003).
36. Aizenberg, J., Rogers, J. A., Paul, K. E. & Whitesides, G. M. Imaging profiles of
light intensity in the near field: applications to phase-shift photolithography.
Appl. Opt. 37, 2145–2152 (1998).
37. Tanaka, T. et al. A novel optical lithography technique using the phase-shifter
fringe. Jpn J. Appl. Phys. 30, 1131–1136 (1991).
38. Hill, K. O., Malo, B., Bilodeau, F., Johnson, D. C. & Albert, J. Bragg gratings
fabricated in monomode photosensitive optical fiber by UV exposure through a
phase mask. Appl. Phys. Lett. 62, 1035 (1993).
39. Vörös, J. et al. Optical grating coupler biosensors. Biomaterials 23,
3699–3710 (2002).
40. Panza, F. et al. Amyloid-based immunotherapy for Alzheimer’s disease in the
time of prevention trials: the way forward. Expert Rev. Clin. Immunol. 10,
405–419 (2014).
41. Vaisocherová, H. et al. Ultralow fouling and functionalizable surface chemistry
based on a zwitterionic polymer enabling sensitive and specific protein detection
in undiluted blood plasma. Anal. Chem. 80, 7894–7901 (2008).
42. Riedel, T. et al. Hepatitis B plasmonic biosensor for the analysis of clinical serum
samples. Biosens. Bioelectron. 85, 272–279 (2016).
43. Piliarik, M., Bocková, M. & Homola, J. Surface plasmon resonance biosensor for
parallelized detection of protein biomarkers in diluted blood plasma. Biosens.
Bioelectron. 26, 1656–1661 (2010).
44. MacDougall, D. & Crummett, W. B. Guidelines for data acquisition and
data quality evaluation in environmental chemistry. Anal. Chem. 52,
2242–2249 (1980).
45. Armbruster, D. A. & Pry, T. Limit of blank, limit of detection and limit of
quantitation. Clin. Biochem. Rev. 29(Suppl. 1),S49–S52 (2008).
46. Fan, X. et al. Sensitive optical biosensors for unlabeled targets: a review.
Anal. Chim. Acta 620, 8–26 (2008).
47. Kolomenskii, A. A., Gershon, P. D. & Schuessler, H. A. Sensitivity and detection
limit of concentration and adsorption measurements by laser-induced
surface-plasmon resonance. Appl. Opt. 36, 6539–6547 (1997).
48. Horvath, R., Pedersen, H. C., Skivesen, N., Selmeczi, D. & Larsen, N. B.
Monitoring of living cell attachment and spreading using reverse symmetry
waveguide sensing. Appl. Phys. Lett. 86, 071101 (2005).
Acknowledgements
The authors thank V. Guzenko (ETH) and W. Arens (IMT) for technical support in
fabrication of the phase mask, S. Tosatti (SuSoS) and S. Zürcher (SuSoS) for consulting
regarding surface science questions and production of the copolymer, A. Nichtl (Roche
Diagnostics) for support with various reagents, H.P. Herzig (EPFL) and A. Naqavi (EPFL)
for support with numerical simulations and J. Hehl and T. Schwarz (ScopeM/ETH) for
STED support. For designing and fabricating numerous hardware components, the authors
thank T. Kissling, R. Rietmann (Roche) and S. Wheeler (ETH). The authors also thank
A. Lieb for support with ZeptoReader-related issues. The authors thank the following for
discussions on various aspects of the project: M. Hennig, K. Mueller, M. Lauer, A. Rufer,
G. Dernick, M. Marcinowski, M. Essenpreis, M. Hein, O. Gutmann, A. Drechsler,
M. Glauser, N. Milicevic, J. Spinke, A. Maurer, C. Patsch, C. Cusulin, J. Fingerle, R. Staack
and A. Poehler. The authors acknowledge the Roche Postdoc Fellowship (RPF) Program,
ETH Zurich and the NCCR Molecular Systems Engineering for funding.
Author contributions
Experiments were designed by V.G., K.-P.S., D.H., T.L., J.V. and C.F. C.F. performed the
calculations for the phase mask and all other optical components. Molographic
experiments were performed by V.G. A.F. performed the numerical simulations and wrote
the evaluation software with support from J.V. and C.F. All authors read and approved the
manuscript for submission.
Additional information
Supplementary information is available in the online version of the paper.
Reprints and
permissions information is available online at www.nature.com/reprints. Publisher’s note:
Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations. Correspondence and requests for materials should be addressed to
J.V. and C.F.
Competing financial interests
The authors declare no competing financial interests.
NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2017.168 ARTICLES
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology 7
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

Methods
All chemicals were obtained from Sigma-Aldrich in at least analytical grade purity if
not stated otherwise. PhSNPPOC chloroformate was purchased from Orgentis
Chemicals. NHS-PEG
12
-OMe was obtained from Iris Biotech. PBS buffer,
streptavidin-AlexaFluor555 and streptavidin-AlexaFluor488 were obtained from
Thermo Fisher Scientific, and the PBS buffer was dosed with 0.05% Tween20. The
PAA-g-PEG-NH-PhSNPPOC copolymer used as a biocompatible coating was
provided by SuSoS. Horse serum was provided by Roche Diagnostics. Human blood
samples were obtained from Blutspende Zürich. All reactions were conducted in
ultrapure MilliQ water (18.2 MΩcm at 25 °C; Merck Millipore).
Establishing a suitable sensing layer for focal molography. The proper design of
the molographic surface on the Ta
2
O
5
thin-film optical waveguide was evaluated by
optical waveguide lightmode spectroscopy (OWLS) with an OWLS 210 instrument
from MicroVacuum (see Supplementary Section ‘Focal molography compared to
OWLS’). The refractometric OWLS technique is based on thin-film optical
waveguide chips comparable with those used for focal molography. This allows
direct transfer of the OWLS findings to chips designed for molography. OWLS chips
were obtained from MicroVacuum and sputtered with 10 nm Ta
2
O
5
at IMT. All
cleaning and coating reactions were carried out in glassware or Teflon chambers.
Chips were cleaned with toluene followed by isopropanol and MilliQ water, each
step for 10 min in an ultrasonication bath. After blow drying and surface activation
by a 2 min exposure to oxygen plasma (18 W, Harrick Plasma PDC-32G) the chip
was immediately assembled in the flow chamber and flushed with running buffer.
After baseline equilibration overnight, the desired OWLS experiment
was conducted.
Preparation of sensor chips for focal molography. Thin-film optical waveguides
with a 145 nm Ta
2
O
5
layer were obtained from Zeptosens. An optional prewash with
0.1% aqueous Tween20 solution assured removal of adsorbed proteins. Intensive
washing with isopropanol and MilliQ water was used to remove organic
environmental contaminations. Chips were then cleaned with highly oxidizing
Piranha solution (H
2
SO
4
/H
2
O
2
7:3) for 30 min. (Special precautions have to be
taken for Piranha solution, which reacts explosively with organic material.) After
excessive washing with MilliQ water, the chips were blow dried and activated by
2 min of oxygen plasma. After plasma treatment, the chips were immediately
immersed in the PAA-g-PEG-NH-PhSNPPOC graft copolymer coating solution
(0.1 mg ml
–1
in 1 mM HEPES pH 7.3 for 20 min). After layer formation, the coated
chips were cleaned with MilliQ water. To fully passivate the layer, dried chips were
immersed in a 25 mM methyl chloroformate solution in anhydrous acetonitrile
containing 2 equiv. N,N-diisopropylethylamine for 5 min. This step assured that
all residual unbound primary amines from the polymeric backbone were not
interfering with the photolithographic process of amine generation. The coated
chips were washed with ethanol and MilliQ water, and blow dried by a nitrogen jet.
Prepared sensor chips could be stored for several months in the dark at 4 °C until
further use.
Photolithographic process for formation of molograms. Photolithography was
conducted in a yellow-lit room (L58 W/62; Osram) to avoid unwanted
photoactivation. To avoid 390 nm stray light during illumination, most parts of the
illumination unit were fabricated with black non-reflecting surfaces. A copolymer-
coated sensor chip was placed inside the illumination unit (Supplementary Fig. 1).
After alignment of the chip and the phase mask (IMT) the micrometer-sized gap
between chip and phase mask was filled with DMSO containing 0.1% aq.
hydroxylamine. The photolithographic exposure at 390 nm was conducted with a
BlueWave LED Prime UVA light source from DYMAX equipped with a light guide.
Illumination was performed at a distance of 20 cm with the light source at full
intensity (resulting in an averaged irradiation power of ∼4.6 mW cm
–2
). After gently
separating the phase mask and chip and washing with ethanol and MilliQ water, the
activated ridges were functionalized with amine reactive substances such as 1 mM
sNHS-biotin. To minimize interference in the assay, any remaining PhSNPPOC
groups were removed by full chip illumination (flood exposure) for 5 min under
DMSO containing 0.1% hydroxylamine. An optional blocking step of
photoactivated primary amines was either performed as described above with
methyl chloroformate, or with a 1 mM aqueous NHS-PEG solution at pH 8.5.
Confocal laser scanning/STED microscopy. Confocal laser scanning microscopy
images of AlexaFluor488-labelled streptavidin were recorded on a Zeiss LSM 510.
STED microscopy images were taken on a Leica SP 8 STED microscope. Fluorescent
streptavidin (AlexaFluor555) was excited at 555 nm with a white light laser and
depleted at 660 nm by the STED laser. Images were deconvoluted with Huygens
Professional and further processed with ImageJ/Fiji software and Python.
Molographic measurements. Thin-film optical waveguides carrying molograms
were measured with the ZeptoReader49 (Zeptosens) at a wavelength of 635 nm for
the incident and diffracted light, without any optical filter in the detection path. For
more information on the ZeptoReader’s mode of operation, see Supplementary
Fig. 3. The object plane containing the molographic foci (900 µm below the
molographic pattern on the chip) was imaged with the camera objective of the
ZeptoReader. This distance marks the plane on which the three-dimensional voxels
are located, where the diffracted light of all curved mologram lines positively
interferes to a focal spot in space. Typical instrument parameters were as follows:
image integration time of 1 s and a grey filter value of 0.001 in the illumination path
of the ZeptoReader. For real-time evaluation of molographic data, automation
(AutoHotkey) and evaluation (MATLAB) scripts were written.
Sandwich immunoassay (TSH). A precomplex of 100 nM SAv and 200 nM
biotinylated anti-TSH F(ab′)2 in PBS-T buffer (0.05% Tween20) were incubated for
10 min before being immobilized at a [NH-biotin|NH-PEG
12
] mologram. TSH was
injected at a concentration of 20 μgml
–1
in PBS-T followed by formation of the
sandwich complex by 20 μgml
–1
anti-TSH IgG. All reagents were provided by
Roche Diagnostics.
Determining specific antibody affinity in a direct binding assay in serum. An
antibody against the C terminus of the human β-amyloid (1–40) peptide was
obtained from Abcam, and the corresponding biotinylated peptide was supplied by
AnaSpec. A preformed complex of streptavidin and the N-terminal biotinylated
β-amyloid (ratio 100:200 nM) was bound to a [NH-biotin|NH-PEG
12
] mologram.
Varying antibody concentrations ranging from 66.6 to 0.6 nM in PBS-T or 20%
horse serum were injected with a flow of 50 μl min
–1
in PBS-T running buffer
(0.05% Tween 20), followed by regeneration by 10 mM glycine pH 2.0.
Therapeutic antibody detection in human plasma. Precomplex formation of
100 nM SAv and 30 nM biotinylated idiotypic capture antibody provided by Roche
Diagnostics was performed in PBS-T buffer (0.05% Tween20) for at least 10 min.
Immobilization of the complex was achieved on a [NH-biotin|NH-PEG12]
mologram. Running buffer (PBS-T) was supplemented with 300 mM NaCl. Human
plasma prepared from EDTA stabilized blood samples was spiked with a therapeutic
antibody provided by Roche Diagnostics. A concentration series of the plasma
sample ranging from 400 to 0.195 nM of the therapeutic antibody was prepared,
followed by a 1:10 dilution with running buffer before analysis by focal molography
in triplicate. Injections of the individual samples were carried out at 10 µl min
–1
for
20 min. Concentrations were applied in the order high to low for the first series and
randomly for the second and third replicates. A stabilized endpoint signal for LOD
evaluation was achieved by washing at 50 µl min
–1
. Between each analyte injection,
the whole SAv precomplex was renewed by stripping the chip with pure DMSO
followed by 8 M guanidine hydrochloride. To evaluate signal variations for various
samples, lithography and plasma donor effects binding curves were evaluated in
triplicate for each of the three parameters (see Supplementary Table 3).
In situ cell culture measurements. A hybridoma cell culture line secreting
glutathione S-transferase (GST) directed antibodies was provided by Roche
Innovation Center Basel. The progression of antibody production was detected
in situ by focal molography through binding to Protein A, which contains four
high-affinity binding sites capable of interacting with the Fc region of IgG. For this
purpose, we used an open well plate holder carrying four molographic sensor chips
as a bottom plate. One well contained a [NH-biotin/SAv/Bi-Protein A|NH-PEG
12
]
mologram, and a second contained a control surface with a [NH-biotin/SAv|NH-
PEG
12
] pattern. Each of these three positive/negative pairs per chip was covered with
only cell culture medium, 0.5 × 10
6
cells ml
−1
or 2.0 × 10
6
cells ml
−1
. Biotinylated
Protein A was obtained from Thermo Fisher Scientific. Both wells were covered
with cell culture medium, with or without cells, respectively. The cell-chip assembly
with 6 × 4 experiments was monitored inside the ZeptoReader without further
temperature or CO
2
control. Statistical analysis was conducted using two-way
ANOVA. A Pvalue of <0.05 was considered significant. For the statistical evaluation,
at least 67 molograms were used.
Data availability. The data that support the findings of this study are available from
the corresponding author upon reasonable request.
References
49. Pawlak, M. et al. Zeptosens’protein microarrays: a novel high performance
microarray platform for low abundance protein analysis. Proteomics 2,
383–393 (2002).
ARTICLES NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2017.168
NATURE NANOTECHNOLOGY |www.nature.com/naturenanotechnology
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

In the version of this Article originally published, the illumination pattern below the phase mask was incorrectly positioned in Fig. 1b (ii)
and Zeptosens was misspelled in two instances in Methods. ese errors have been corrected in all versions of the Article.
Erratum: Focal molography is a new method for the in situ analysis of molecular
interactions in biological samples
Volker Gatterdam, Andreas Frutiger, Klaus-Peter Stengele, Dieter Heindl, Thomas Lübbers, Janos Vörös and Christof Fattinger
Nature Nanotechnology https://doi.org/10.1038/nnano.2017.168 (2017); published online 25 September 2017; corrected online
25 October 2017.