FULL ARTICLE
Ultramicroscopy: 3D reconstruction
of large microscopical specimens
K. Becker
**
; 1; 2
,N.Ja¨hrling
**
; 1; 2
, E. R. Kramer
3
, F. Schnorrer
4
and H.-U. Dodt
*
; 1; 2
(K. Becker and N. Ja¨hrling contributed equally to this work)
1
Vienna University of Technology, Institute of Solid State Electronics, Dept. of Bioelectronics, Floragasse 7, 1040 Vienna,
Austria
2
Center for Brain Research, Medical University of Vienna, Spitalgasse 4, 1090 Vienna, Austria
3
Max Planck Institute of Neurobiology, Dept. of Molecular Neurobiology, Am Klopferspitz 18, 82152 Martinsried,
Germany
4
Research Institute of Molecular Pathology (IMP), Dr. Bohrgasse 7, 1030 Vienna, Austria
Received 11 October 2007, revised 13 November 2007, accepted 15 November 2007
Published online 9 January 2008
Œ
Supporting information is available on the WWW under www.biophotonics-journal.org
Key words: microscopy, ultramicroscopy, imaging, lightsheet illumination, mouse embryo, Drosophila
PACS: 07.05.Pj, 87.64.-t
#
2008 by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1. Introduction
Various new microscopical techniques, like confocal
and 2-photon microscopy have been developed in
recent decades [1]. Although both techniques allow
excellent spatial resolution, they suffer from limited
fields of view and penetration depths of less than
1 mm. Thus, confocal and 2-photon microscopy can-
#
2008 by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Journal of
BIOPHOTONICS
Ultramicroscopy is a microscopical technique that allows
optical sectioning and 3D reconstruction of biological
and medical specimens. While in confocal microscopy
specimen size is limited to several hundred micrometers
at best, using ultramicroscopy even centimeter sized ob-
jects like whole mouse embryos can be reconstructed
with micrometer resolution. This is possible by using a
combination of a clearing procedure and the principle of
lightsheet illumination. We present ultramicroscopic 3D
reconstructions of whole immunohistochemically la-
belled mouse embryos and adult Drosophila, giving de-
tailed insight into their anatomy. Its speed and simplicity
makes ultramicroscopy ideally suited for high-throughput
phenotype screening of transgenic mice and thus will
benefit the investigation of disease models. Image obtained from a mouse embryo E12.5, recon-
structed from 4
413 optical sections
*
Corresponding author: e-mail dodt@meduniwien.ac.at, Phone: +00 431 427 762 806. Fax: +004 315 880 136 299
**
e-mail: klaus.becker@meduniwien.ac.at, nina.jaehrling@meduniwien.ac.at, Phone: +00 431 427 762 807. FAX: +004 315 880 136 299
J. Biophoton. 1, No. 1, 36–42 (2008) / DOI 10.1002/jbio.200710011

not image whole objects larger than a few hundred
micrometers [1]. Furthermore, confocal microscopy
is not possible with objectives of NA about 0.1 for
large fields of view. Due to these limitations histolo-
gical sections are still the standard way to obtain 3D
reconstructions from large specimen like whole
mouse embryos. However, obtaining good recon-
structions by mechanical slicing is a laborious proce-
dure often with limited success. For a reliable 3D-re-
construction up to several thousand slices have to be
spatially aligned and although in recent years com-
putational algorithms have been developed to par-
tially automate this process, their results are often
unsatisfactory. The reason is that mechanical distor-
tions due to the slicing procedure unavoidably cause
artefacts, which make a correct alignment often im-
possible [2, 3].
To overcome these drawbacks we developed a
microscopical technique that enables us to substitute
mechanical slicing by optical sectioning also for large
specimens [4]. For our technique we use the 100-
year-old idea of light sheet illumination, once termed
ultramicroscopy [5]. This approach has also been
used in other approaches like OPFOS and SPIM [6,
7]. In ultramicroscopy, the specimen is illuminated
from the side by a thin lightsheet generated by one
or more cylinder lenses (Figure 1). If this lightsheet
is made sufficiently thin, all parts of the specimen
above or below the focal plane are not illuminated
and thus no out of focus light is generated [4]. This
straightforward approach further bears the advan-
tage that no light from out-of-focus planes has to be
excluded later by a pinhole as in confocal microscopy.
So, no superfluous light from out-of-focus planes con-
tributes to photobleaching of the specimen.
In contrast to confocal microscopy, illumination
and observation pathways are separated in ultrami-
croscopy, and therefore also low numerical aperture
(NA) objectives can be used. These objectives pro-
vide large fields of view.
Since in ultramicroscopy a thin lightsheet traverses
the specimen perpendicular to the objective, speci-
mens have to be sufficiently transparent. Because
most interesting biological and medical specimen, like
whole mouse embryos, are opaque, we applied a spe-
cial clearing technique that, like ultramicroscopy, was
invented at the end of the 19
th
century [8]. The princi-
ple of this clearing technique depends on imbuing the
specimen with a medium having the same refractive
index as protein. Thus, the refractive indices of the in-
tra- and extracellular compartments of the specimen
become equal, and light can traverse the specimen un-
hindered without scattering. If light absorption by the
specimen is not high the specimen appears transpar-
ent. By carefully applying this clearing technique we
have been able to clear complete rat embryos up to
2 cm in size and entire adult Drosophila.
2. Experimental
2.1 Specimen preparation
Mouse embryos: We obtained E12.5 mouse embryos
from wild-type mice, sacrificed by cervical disloca-
tion. The procedures were approved by the local an-
imal-care committees.
Whole-mount immunohistochemistry: Immuno-
histochemical labelling was carried out by using sec-
ondary fluorescent antibodies. Embryos were fixed
overnight in DENT’s fix (1 part DMSO and 4 parts
methanol) and bleached overnight in 1 part 30%
H
2
O
2
and 2 parts DENT’s fix [9]. Afterwards, they
were washed three times in TBS for 30 min and incu-
bated for two days at room temperature in the primary
antibody solution (monoclonal antineurofilament 160
clone NN18 from Sigma-Aldrich, Germany, diluted
1 : 200 in blocking serum containing 4 parts calf serum
and 1 part DMSO). Then, the embryos were washed
three times for one hour in TBS and incubated for two
days at room temperature in the secondary antibody
solution (goat antimouse conjugated to Alexa488
from Invitrogen, USA, diluted 1 : 200 in blocking ser-
um). Afterwards, we washed the embryos at least five
times in TBS (one hour each). The embryos E12.5
were retained and dehydrated in ethanol (50%, 70%,
80%, 96%, 3
100%, each concentration for one
hour, last step overnight). Finally we transferred the
embryos into the clearing solution containing 2 parts
benzyl benzoate (Sigma-Aldrich, Germany) and one
part benzyl alcohol (Sigma-Aldrich, Germany)
(BABB, refractive index 1.559) (Figure 2). The em-
bryos were left in the clearing solution, which was
changed three times, for at least two days.
Drosophila: White-eyed adult Drosophila males
(w
1118
) were anaesthetised by ether and transferred
Figure 1 Principle of ultramicroscopy. The sample is illu-
minated by a blue laser forming a thin sheet of light.
Fluorescent light is thus emitted only from a thin optical
section and collected by the objective lens.
J. Biophoton. 1, No. 1 (2008) 37
FULL
ARTICLE
#
2008 by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheimwww.biophotonics-journal.org