Donna Arndt-Jovin : Research Interests and Projects

Back to Donna Arndt-Jovin

Research Interests

Signal transduction processes of the erbB tyrosine kinase receptor family in living tissue culture cells and tumor tissue.
Development of new fluorescence imaging modalities for rapid, in vivo cell and organism imaging.
Chromatin structure and function in vivo,

Several research projects of the group are summarized.

Signal Transduction:
Reaching out for growth factor signals

Activation of the epidermal growth factor receptor (EGFR, erbB1) and related family members (erbB2, erbB3, erbB4) by binding of peptide effector ligands initiates signaling cascades controlling numerous cellular processes such as DNA replication, division, migration and differentiation. Phosphorylation of particular residues leads to binding of specific adaptor proteins channeling into various parallel signaling pathways. We discovered a new phenomenon involved in growth factor-mediated signal transduction using prolonged live cell imaging with epidermal growth factor coupled to Quantum Dots (QD-EGF) and single molecule sensitivitive detection in optical sectioning microscopes such as the PAM (see related section).

We observed that the activated transmembrane EGFR undergoes retrograde transport on cellular extensions known as filopodia. The receptor-ligand complexes attach to actin filaments constituting the core of the filopodia, ‘treadmilling’ together with these cytoskeletal elements towards the cell body, where they undergo endocytosis. (Fig.1).

View larger version: [In this window] [In a new window]

Fig. 1: Simultaneous tracking of GFP-actin and EGF-QD-erbB1. Selected frames from time series of EGF-QD-erbB1 (red) undergoing transport on the filopodia of HeLa cells expressing GFP-actin (green). After bleaching of GFP-actin (in yellow box), the QDs (white arrows) are transported to the cell body (upper left corner) at the same rate as the unbleached GFP-actin. Right panel: DIC image of filopodia. Images are single 1 μm confocal sections (from Lidke et al. 2005).

We have proposed that the filopodia and the directed transport serve to detect the presence and gradients of effector molecules far from the cell body, thereby mediating appropriate cellular responses. We subsequently determined the diffusion constant of unactivated QD-bound EGFR with high precision, demonstrating the absence of comparmentalisation of the receptors on filopodia, which rather are in equilibrium with molecules on the remainder of the plasma membrane encompassing the cell body.

A key question regarding the process of retrograde transport is how the activated receptors attach to the actin filaments (Fig. 2 scheme). Experiments based on siRNA and dominant negative mutants as well as knockout mouse fibroblasts have ruled out the involvement of motor proteins such as myosin VI as well as direct linkage through EGFR. Recently we made single and multiple mutants in four key tyrosine moieties in the cytoplasmic C-terminus of EGFR that are the loci of specific phosphorylation serving for recognition by the adaptor proteins such as SHC, GRB2, and CBL and can show inhibition of retrograde transport.

Fig. 2. Schematic depiction of filopodial retrograde transport of EGFR.The unactivated, unliganded EGF receptor is shown as monomers in the cell membrane. Binding of EGF or EGF-QD (green and red objects) leads to conformational rearrangements in ectodomain II (including extension of the dimerization loop) and other domains (black cytoplasmic region), potentiating the stabilization of an “active dimer” competent for auto- and transphosphorylation. The phosphoprotein (-P) interaction mediated by adapter proteins (blue rectangles) leads to a physical linkage of the receptor complex to F-actin filaments (black lines) and a shift from restricted diffusion (pair of opposed arrows) to directed retrograde transport towards the cell body where the receptor complex is endocytosed through clathrin coated pits.

Nanoparticle diagnostics, detection of glioblastoma cells in human brain biopsies

The project is to create improved diagnostics for human erbB over-expressing tumors.
In collaboration with neurosurgeons at the University of Göttingen Medical School we have produced specifically conjugated quantum dots that targeted Her1 expression which is upregulated on greater than 30% of all human gliomal brain tumors. Both Mab- and EGF- (epidermal growth factor) coupled QDs specifically recognized not only high-grade glioma biopsies but also were able to distinguish tumor cells in low-grade biospies where no gadolinium positive signal can be visualized. Excellent delineation of tumor tissue from normal brain in fresh biopsies was achieved even in the case of low-grade tumors.

A. T-1weighted MRI brain scan with no gadolinium positive signal of a grade II oligodendroma. (B-E) Digital macrophotographic images of ex vivo stained biopsies from the resected tumor stained with targeted (B-D) QD probes or (E) untargeted QDs taken with the same magnification and the same exposure times.

The results of the high grade tumor protocol was published in 2009 in IEEE Transactions in Nanobioscience and the protocol for low-grade tumors was published in PLoS ONE in June 2010, http://dx.plos.org/10.1371/journal.pone.0011323. The success of the protocol has allowed the clinical collaborators to propose a Phase 1 assessment of this diagnostic for patients in the near future. The clinical application of the diagnostics developed in the project will benefit the citizens of the EU by helping to raise life-expectancy for patients with glioblastoma brain tumors.

We have also been able to target both resistant and sensitive Her2 expressing breast tumor cells with specific affibody-conjugated probes formulated in collaboration with the group of Manfred Konrad, MPIBPC. A conjugated prodrug converting enzyme and affibody construct were internalized by the Her2 expressing tumors. In order to increase the level of enzyme available for prodrug conversion a microcapsule technique is on-going and the goal is to produce a vehicle that would target the Her2 cells and make them more sensitive to chemotherapy.

The Programmable Array Microscope

The Laboratory of Cellular Dynamics has developed a high-speed, optical sectioning microscope with single molecule sensitivity called the Programmable Array Microscope (PAM)

This project is headed by Thomas Jovin.

The new Generation 3 PAM design consists of the following elements:

1. Use of a TI DMD digital micromirror device as the SLM control element with Visitech control electronics and software
2. Telecentric relay between microscope and DMD.
3. DMD condensor from DMD to secondary afocal relay.
4. Afocal telecentric relay from DMD to CCD.
5. Image combiner and CCD relay.

Since the PAM microscope utilizes the SLM in both the excitation and emission pathways the optical path presents unique problems for maintaining high resolution as well as achromatic aberration-free imaging. In addition, since the DMD functions by deflecting the “on” and “off” pixels at an angle of +/- 12º, the imaging beam will have an angle of 24º to the optical axis. A unique solution to this optical problem was achieved by the scientists of the LCD and this optical design has been protected by patent applications, EP 1000 3066.7, USP 61/316,671. Careful selection of glass for the optical components has assured that the final demonstrator has the best properties for imaging with single molecule sensitivity. A demonstrator microscope is nearing completion. A European SME has signed a ‘Letter of Intent’ and is building a second demonstrator with the commitment to start commercial production of the PAM3 in late 2010/early 2011.

PAM v3 demonstrator - covers and baffles removed to show optical components more clearly; optical paths drawn in green.

New software capabilities have been generated for the PAM. Real-time imaging control software was developed as well as software to interleave multiple light sources and to take multiple emission wavelength images. Pattern interleaving was developed as well. Recently a new n-dimensional image rendering software package to visualize the often four- or five-dimensional data generated by the PAM was realized. This software is able to show arbitrary projections of large (multi-GB) data sets interactively. In addition, it can perform spectral rendering for the images taken for full spectrum PAM imaging. Arbitrary combinations of linked XY/XZ/YZ or XY/XZ/XC projections are possible. Other capabilities include length measurements and histogram analysis.

We have demonstrated single-molecule sensitivity with 16 ms acquisition using laser light sources for 1k x 1k pixel images and faster acquisition for regions of interest (ROIs). Lifetime imaging with the PAM was realized with incorporation of a phase modulated light source and camera system in collaboration with Lambert Instruments.

Several more new capabilities have been realized for PAM imaging. The goal of high-speed imaging in live cells is often to collect data sets over long time periods in order to determine the effects of drugs, inhibitors or stimulants on the cells. Such studies have usually resorted to non-confocal methods in order to reduce bleaching and detrimental effects of high excitation intensities. The PAM offers a unique capability for such measurements since the instrument itself can regulate the light dosage so as to avoid bleaching and toxicity. The MLE-PAM (minimized light exposure PAM) is described in detail in the report for WP11. In short, this software allows imaging of living embryos and cells over periods of hours instead of 10s of minutes. In addition, because the imaging is under computer control, cells or sub-cellular compartments that are moving over the time of the experiment can be tracked and kept in focus without operator intervention. These new capabilities of the PAM have been recently published in an article in the J. of Microscopy, http://www3.interscience.wiley.com/journal/123548346/abstract. The PAM has been used for imaging of tumor biopsies and demonstrated improved capabilities for distinguishing localized tumor cells with 3-D imaging compared with other systems. (See the images above). These data were recently published in PLoS ONE and suggest that for a number of applications in cell biology the commercialization of the PAM could include a suite of specialized instruments as well as the highly versatile PAM for the basic research market. Instruments could be targeted for the health care market for specific diagnostic tests, such as an operating theater PAM.

Transcriptional control in development:
Polycomb-group (Pc-group) protein repression

Both larval and adult body segment identities are determined early in embryogenesis and maintained through up to ten cell divisions before differentiation by an exquisite control of the temporal and spatial expression of the homeotic genes. The original pattern of repression is set up by the segmentation and gap genes. As the products of these genes disappear the Polycomb group (PcG) genes take over the task of maintaining repression of the homeotic genes in the proper tissues.

We are interested in the mechanism by which PcG genes repress transcription and maintain silencing during development. We have analyzed qualitatively and quantitatively the subcellular 3-dimensional distribution of three PcG proteins, Polycomb, Polyhomeotic and Posterior sex combs, in fixed whole-mount Drosophila embryos by multicolor confocal fluorescence microscopy. Our observations contradict several widely held views about the mechanism of PcG silencing. In particular, we have found that the repression complex does not remain on the chromatin through mitosis but is reassembled after each cell cycle. Questions about the marking of these assembly sites as well as the nature and composition of the repression complex are under investigation.

Determination of the lifetime and stability of Polycomb group protein repression complexes in living Drosophila

We use fluorescence recovery after photobleaching (FRAP) microscopy to determine the kinetic properties of Polycomb group (PcG) proteins in whole living Drosophila organisms (embryos) and tissues (wing imaginal discs and salivary glands). Green fluorescent protein (GFP) fusion proteins of two PcG proteins, polycomb (PC) and polyhomeotic (PH), have been constructed and transgenic fly lines established (Fig. 1).

View larger version: [In this window] [In a new window]

[In this window] [In a new window]

Fig.1: Confocal images of PcG-GFP expressing Drosophila embryos and larval tissues. A-D (left image): A,B, embryo expressing PhGFP (UAS promoter, engrailed driver); C,D, embryo expressing PcGFP (Pc endogenous promoter). E-I (right image): E,F, wing imaginal disc PhGFP (UAS promoter, Bx driver, wing blade); H,I, wing imaginal disc PcGFP.

Polycomb group (PcG) genes are essential genes in higher eukaryotes responsible for the maintenance of the spatially distinct repression of developmentally important regulators such as the homeotic genes. Their absence, as well as overexpression, causes transformations in the axial organization of the body. Although protein complexes have been isolated in vitro little is known about their stability or exact mechanism of repression in vivo.

FRAP microscopy has been used to determine biophysical constants for protein binding in cell nuclei in other systems but a formalism for treatment of discrete complexes has been lacking. We have developed computer simulations with new models for spatially distributed protein complexes in systems showing both diffusion and binding equilibria and the results are compared with experimental data. (manuscript in preparation).

The nuclear contents of whole live Drosophila embryos and larval tissue are very mobile. To determine dissociation constants we developed 3D-inverseFRAP, a 4D-imaging photobleaching/recovery technique that consists of bleaching the whole nucleus except for a small region surrounding a fluorescent locus of interest, recording a time series of confocal z sections for several hundred seconds after bleaching, tracking the fluorescent locus in 3 dimensions (3-D), alignment of the locus in 3-D and calculation of the spot intensity using a weighted region of interest (Figs 2 & 3).

View larger version: [In this window] [In a new window]

Fig. 3: 3D-iFRAP.(A) Overview of part of a wing disc with cross-hair on the fluorescent locus selected for iFRAP. Fluorescence depletion was calculated from a recorded series of Z-stack images over time after bleaching and image registration. (B) Magnified XY image of the nucleus before bleaching. Cross-hair set in the bleach region next to the unbleached locus. The images have been registered to correct for the movement of the chromatin and the nucleus itself and the fluorescent locus is centered in the final analysis image time stack. (C and D) Kymographs. X-time and time-Y views, respectively of the registered images at the slices in the XY image corresponding to the cross-hair in B. In (C) dissociation of the fluorescent molecules from the spared locus can be seen over time. In (D) a bleached region is shown over time. Note the transition from before bleaching to after bleaching indicated by the arrows in the time planes.

In embryos we have found dissociation rate constants of k_off= 0.047/s for PHGFP and k_off= 0.029/s for PCGFP. In larval tissue complexes we found dissociation to be somewhat slower but fully exchangeable in less than 6 minutes (Fig 4). These data rule out a model for repression in which PcG complexes would mask the chromatin and make it inaccessible to transcription factors or other proteins.

View larger version: [In this window] [In a new window]

Fig. 4: Dissociation rate constants for PhGFP. Individual PcG protein loci in the diploid nuclei of the embryos and wing imaginal discs were subjected to 3D-iFRAP. Fluorescence decay curves were fitted to a single exponential function. Histogram of dissociation rate constant of individual PhGFP loci obtained from 3D-iFRAP experiments in embryos and larval wing imaginal discs are show in red and black, respectively.

Salivary glands have polytene nuclei where the complexes bind to the gene loci on the aligned chromatids. By multicomponent analysis, we were able to analyze complexes on individual polytene chromosome bands in intact glands. Reproducible individual dissociation rates (Fig 5) (though similar within a factor of 5) presumably reflect the specific mixture of PcG and non-PcG auxiliary proteins at individual PREs (Fig 6). Our data demonstrate that the mechanism of long-term silencing maintained over 10 cell generations is achieved by mass action equilibria of freely dissociable complexes undergoing continuous exchange and thus available for epigenetic reprogramming by changes in the local concentration of proteins that compete for the same binding sites [ref: Ficz et al. Development 132, 3963-3976 (2005)].

View larger version: [In this window] [In a new window]

Fig. 5: FRAP curves for individual bands in the larval salivary gland nuclei. Selected confocal fluorescence images from of part of a whole salivary gland nuclei undergoing a bleach and recovery time sequence. Selected time points, first being prebleach image and 6th being final image in the series. (A) Upper panels, overview of part of nucleus with bleach box indicated in 2nd image. Lower panels, magnified region used for analysis. (B) FRAP curves for sequential bleaching of a single PcGFP locus showing reproducible reassociation kinetics.

View larger version: [In this window] [In a new window]

Fig. 6. Dissociation rate constants for PhGFP and PcGFP. Dissociation rate constants for PhGFP (blue) and PcGFP (red) plotted against the normalized concentration of binding sites for individual loci.

Study of nuclear architecture: Homologous pairing and transvection

Chromatin structure has been resolved at the nucleosomal level, but the higher levels of organization of the interphase chromosomes as well as the arrangement of functional elements within the nucleus are still largely undefined and their structural features disputed. There are insufficient experimental data available at present to form a coherent picture of the nuclear architecture which takes into account the broad range of specialized functions found in the various nuclei of the multicelluar eukaryotic organism and their highly dynamic nature in response to metabolic and developmental changes in their environments. Our present project is to understand the physical basis for chromosome recognition and pairing and the extent to which this occurs in diploid interphase nuclei. The questions we are addressing are: how flexible are interphase chromosomes, how do mutants in various genes affect the onset or extent of pairing, and what are the forces involved in chromosome movement?

In this context we have recently investigated the physical basis for transvection. Transvection is the phenomenon by which the expression of a gene can be controlled by its homologous counterpart in trans, presumably due to pairing of alleles in diploid interphase cells. The phenomenon was first described for genes of the Bithorax complex (BX-C) and it is documented by a large body of genetic data for this locus. We have tested transvection models which postulate direct pairing and/or looping of the DNA strands compared to those which postulate transport of macromolecules from juxtaposed loci. Our data were able to eliminate the latter model. We have determined the proximity of homologous chromosomes as a function of embryo and tissue development for these genes and are extending these studies to other loci.

Our working model is a multipoint recognition of sequences dispersed along the chromosome or "cardigan model" with each site having a finite binding constant and equilibria between pairing and dissociation.