PNAs were synthesized and purified in parallel in microwell plates for a series of target sequences (for details see [18]). This process has advantages in those cases where small amounts of probe material but a high diversity of probe types under high quality and purity conditions are required. For the experiments presented in this article here, we tested several samples of the same purine–pyrimidine mixed PNA sequence with respect to sensitivity and labeling quality of different dyes. MALDI-TOF analysis of the purified labeled products revealed high quality results. The PNA sequences with the respective fluorochromes are listed in Table 1. Based on the concentration value, the probe volume for COMBO-FISH per slide was determined.

For the investigation described here, new type PNA oligonucleotide probes were used which were synthesized and purified according to new protocols. In order to test the hybridization quality of the PNA COMBO-FISH experiments, the three probe types were separately hybridized on methanol/acetic acid fixed peripheral blood lymphocyte cell nuclei and metaphase spreads. In all cases, the cell nuclei showed two fluorescent spots and the centromeres of two middle sized metaphase chromosomes were labeled (data not shown). To better verify the binding specificity and efficiency, we also performed simultaneous two color experiments with different spectrally labeled centromere 9 PNA probe types (one with OregonGreen 488^®, the other with TexasRed^®–Invitrogen, Carlsbad, CA, USA). Merging the green and red color image plane of the same nuclei or metaphases revealed a correct colocalization of the green and red labeling spot indicating competitive binding at the same target site (Figure 1). This seriously supports the centromeric specificity of the probe and reproducibility of the preparation procedure.

In the next step, it was verified that the labeled chromosome is really chromosome 9. For this purpose, the lymphocyte specimens were subjected to another two-color experiment (Figure 2) using the repetitive centromere 9 PNA COMBO-FISH probe labeled with OregonGreen 488^® and a commercially available 9q subtelomere (9qtel) DNA standard FISH probe (MPbiomedicals, Montreal, Quebec, Canada) labeled with rhodamine. In these experiments, the metaphase spreads showed two chromosomes with red spots indicating the DNA probe marking the subtelomere ends of the long arms of human chromosome 9. The same chromosome carries the green signals of the PNA probes in the centromere region. This demonstrates the specificity of our synthesized PNAs applied here for chromosome 9.

By illuminating the sample with a high laser excitation intensity (ten to several hundred kW/cm² Alexa 488^® shows the characteristics of a PHYMOD (Physically Modifiable) dye, i.e., a certain amount of molecules can be transferred to a reversibly bleached state (M_fl → M_rbl). Stochastic recovery of these molecules to the fluorescent state (M_fl ← M_rbl) allows for optical isolation of the detected single molecule signals during acquisition of a time stack of several hundred images of the same specimen region. Therefore we tested PNA COMBO-FISH labeling with Alexa 488^® for principle applications of SPDM using formaldehyde fixed, human mammary epithelial cells of the line AG11132 growing directly on coverslips (Figure 3).

High sensitive SPDM of cell nuclei after COMBO-FISH revealed localization images in the green image plane with PNA probe signals and unspecific background presumably due to label-free fluorescence [19] or some unspecific probe attachment (Figure 3A). The effective optical resolution of this image was given by the density of detected signals and the localization accuracy. Therefore, the frequencies of localization accuracy values were determined for all signals registered within the centromere clusters and the background region (Figure 4). The majority of the signals referring to the centromere clusters revealed a localization accuracy better than 30 nm whereas the majority of background signals showed a worse localization accuracy. Thus, the centromere clusters were discriminated by a localization accuracy threshold of 30 nm (Figure 3B). Moreover it was shown that within these centromere clusters most signals had more than 40 neighboring signals within a circle of 200 nm radius (Figure 3C).

For the experiments, two types of cells were used: (a) peripheral blood lymphocytes from a healthy donor; (b) cells of the line AG11132 (Coriell Institute, Camden, USA), human mammary epithelial cells of a healthy donor which have been established from normal tissue obtained at reduction mammoplasty.

Peripheral blood lymphocytes were isolated using a Ficoll gradient. 20 mL Ficoll (GE Healthcare, Munich, Germany) preheated to 37 °C were slowly covered with 30 mL heparinized whole blood. By means of centrifugation (200 g, 20 min) the cells were separated into four different gradients. The layer with the peripheral blood mononuclear cells was transferred into RPMI 1640/Fetal calf serum (FCS) (Invitrogen, Carlsbad, CA, USA/Biochrom, Berlin, Germany) (1:1, 37 °C). After centrifugation (200 g, 10 min) the supernatant was discarded. The pellet was washed once in 1 × PBS. The lymphocytes were further cultivated in preheated chromosome medium B (Biochrom, Berlin, Germany), which contains phytohaemagglutinin (PHA). After incubation for 72 h at 37 °C in a CO₂-incubator, the medium was spiked with Colcemid (concentration 10 μg/mL) (Sigma-Aldrich, Hamburg, Germany) and the cell suspension was again incubated for 10–30 min at 37 °C. After centrifugation (200 g, 10 min) the supernatant was discarded. Then the hypotonic treatment followed with pre-warmed 75 mM KCl (Sigma-Aldrich, Hamburg, Germany). After incubation for 15 min at 37 °C and another centrifugation (200 g, 10 min) the supernatant was removed. The cells were then fixed by adding slowly −20 °C cold methanol/acetic acid (3:1) while the tube was agitated slightly. This fixation step was repeated 3–4 times.

AG11132 cells were grown in mammary epithelial growth medium (MEGM) supplemented with 4 μL/mL bovine pituitary extract, 5 μg/mL insulin, 5 ng/mL epidermal growth factor, 0.5 μg/mL hydrocortisone (MEGM BulletKit from Lonza, Walkersville, USA), 10⁻⁵ M isoproterenol (Sigma-Aldrich, Hamburg, Germany), 5 μg/mL transferring (Sigma-Aldrich, Hamburg, Germany), 10 mM HEPES buffer (Sigma-Aldrich, Hamburg, Germany) in a standard CO₂-incubator. All cells were seeded onto glass slides (Menzel-Gläser, Braunschweig, Germany) and allowed to attach and grown overnight. Cells were fixed with 4% formaldehyde in PBS when they reached 80% of confluency. At all following steps (see below) drying of the formaldehyde fixed cells was avoided.

The basics of database searching for appropriate combinations of oligonucleotides have been described in detail elsewhere [7]. Due to the completion of the human genome sequencing and the unification of the NCBI chromosome data files, the scope of search has been extended and the algorithms have been improved considerably. Using the annotation information for the region to be labeled from the corresponding NCBI contig files, the nucleotide sequences mapping these regions are identified and all homopurine and homoyrimidine stretches with a minimum length of 15 nucleotides are extracted by a straightforward search algorithm. This list of candidate individual target sequences is inspected for unfavorable stretches, e.g., with largely differing melting points or with periodicities, and for long sequences (more than 30 bases) which can be cut down to one or more shorter ones being more effectively synthesizable. The resulting ‘basic’ set is further processed to eliminate those oligonucleotide probes that form clusters of more than five stretches within any 250 kb genomic region elsewhere other in the genome than the selected target regions. For this purpose, the occurrence within the whole genome of all members of the basic set is listed. We prefer exact search to BLAST-search for three reasons: Firstly, gaps do not play any role at all, because the sequences have to fit exactly for binding. Secondly, we have to know securely each single occurrence of any stretch. And thirdly, nucleotide mismatches would lower the binding affinity, hence rendering non-exact matches to subordinate importance. It should be emphasized that, as purines pair to pyrimidines, it is necessary and sufficient to list both directions of one type of homosequences from the annotated strand. The finite automaton procedures for exact search are programmed in C. In the last semi-automated step, those candidate sequences, which occur frequently in clusters with more than five stretches within 250 kb, are iteratively removed until no further clusters except the labeling sites are left. For the proof of concept and feasibility studies presented here, the search was modified to purine-pyrimidine mixed probes. Here, an oligonucleotide (AAT CAA CCC GAG TGC AAT) has been chosen that uniquely exists as a repetitive sequence in the centromere region of chromosome 9 (see also [20]).

Only small quantities but high quality and purity of PNA sequences were needed. Therefore the PNA probes carrying different types of dye molecules (OregonGreen 488^®, TexasRed^®, Alexa 488^®) were synthesized in 384 microwell plates followed by a low-cost parallel purification. Purity and recovery of the PNAs were determined by MALDI-TOF mass spectrometry and HPLC. Details will be published elsewhere [18].

10–15 μL of the lymphocyte-solution was dropped on a wet slide. Slides were then air-dried. The pre-treatment of methanol/acetic acid fixed cells included multiple steps such as 0.7% Triton X-100/0.1% Saponin (Merck, Darmstadt, Germany/Serva, Heidelberg, Germany) in 2 × SSC for 30 min at room temperature, RNase A digestion (Sigma-Aldrich, Hamburg, Germany) (100 μg/mL in 10 mM Tris HCl, 1 h at 37 °C), pepsin-treatment (Sigma-Aldrich, Hamburg, Germany) (10% pepsin with a concentration of 100 mg/mL was diluted in a 0.01 M HCl, pH 2.3), dehydration of specimen in an aqueous ethanol series of 70%, 90%, and 100%. Denaturation was carried out in pre-warmed 70% formamide (AppliChem, Darmstadt, Germany) in 2 × SSC (pH 7.0) for 5 min at 75 °C. After denaturation, the slides were transferred back into an ice-cold ethanol series. Slides were air-dried.

The pre-treatment of formaldehyde fixed AG11132 cells included multiple steps such as 0.5% Triton X-100 in PBS, 20% glycerol (Sigma-Aldrich, Hamburg, Germany) in PBS, repeated freezing-thawing in liquid nitrogen and incubation in 0.1 N HCl. Slides were subsequently stored in 50% formamide in 2 × SSC (pH 7.0) at 4 °C until usage. Denaturation was carried out in pre-warmed 70% formamide in 2 × SSC (pH 7.0) for 3 min at 72 °C. After denaturation, the slides were transferred back into the 50% formamide in 2 × SSC (pH 7.0). The following steps (hybridization, washing steps and counterstaining) were accomplished equally for both preparation methods.

For hybridization, about 200 pmol/μL of the fluorescence labeled centromere 9 PNA probe (Table 1) was added to 20 μL hybridization buffer (0.6 M MgCl₂, 3 M NaCl, 0.5 M sodium acetate, pH 7.0; Sigma-Aldrich, Hamburg, Germany). The mixture was applied onto the slide and covered with a cover slip, carefully sealed with rubber cement. Hybridization was done overnight at 37 °C in a humidity chamber. The next day, the rubber cement was removed and the slides were washed in 2 × SSC for 5 min at 42 °C. This step was repeated twice followed by another stringent washing step in 0.5 × SSC at 60 °C. The specimen was counterstained with TOPRO-3-Iodide (Invitrogen, Carlsbad, CA, USA) (1:1000) for 45 s, followed by washing three times in 1 × PBS and embedding in ProLong^® Gold antifade reagent (Invitrogen, Carlsbad, CA, USA). Finally the slides were covered with a coverslip (no. 1.5, 170 mm thickness, Menzel-Gläser, Braunschweig, Germany) and sealed with rubber cement (Marabu, Tamm, Germany) for microscopy.

For the acquisition of high resolution fluorescence images, an Ultra-View confocal spinning disc (PerkinElmer Life Science) on a Nikon TE2000-E inverted microscope was used (Nikon Imaging Center, Bioquant, Heidelberg), which was equipped with a Plan Apo 100x/NA1.4 oil immersion objective lens and lasers for excitation wavelengths of 488, 568 and 647 nm. The fluorescence was detected via appropriate band pass filters on an electron-multiplying charged-coupled device camera (EM-CCD, C9100-50, Hamamatsu). Emission of OregonGreen 488^® was detected via band pass filter 527 (55) nm, TexasRed^® via 625 (70) nm emission filter and TOPRO-3-iodide via 705 (90) nm. The lateral scanning area was up to 1000 × 1000 pixels corresponding to a voxel size laterally of 82 nm × 82 nm. Within this scanning area, a ROI (region of interest) was selected according to the size of the cell nucleus recorded. The step width for optical sectioning was 200 nm in axial direction. All the 3D image stacks of the cell nuclei were evaluated by Matlab based algorithms.

Localization measurements were performed with the SPDM setup [4]. Optical isolation of signals was achieved utilizing a light induced reversibly bleached state [21–23] of Alexa 488^®. By illuminating the sample with a laser excitation intensity of 10 kW/cm² to several 100 kW/cm² some fluorophores are bleached irreversibly (M_fl → M_ibl), but another amount is transferred to a reversibly bleached state (M_fl → M_rbl). Stochastic recovery of these molecules to the fluorescent state (M_fl ← M_rbl) can be used for optical isolation of the detected single molecule signals. This method allows the usage of a series of conventional fluorophores (PHYMOD dyes) for localization microscopy.

For the present experiments, a DPSS laser with a wavelength of 488 nm (Sapphire HP 488, Coherent, Dieburg, Germany) was used. The laser beam was expanded by a factor of 2.5 before being focused into the back focal plane of an oil immersion objective lens (HCX PL APO, 63x, NA = 0.7–1.4, Leica, Wetzlar, Germany). Fluorescent light emitted by the fluorophores in the sample passed a dichroic mirror (AHF Analysetechnik, Tübingen, Germany) and a blocking filter (F73-491, AHF Analysentechnik, Tübingen, Germany) before being focused onto the CCD chip of a very sensitive camera (SensiCam QE, PCO Imaging, Kehlheim, Germany). An additional lens can be mounted in the excitation pathway for increasing the laser intensity in the object plane to obtain appropriate conditions for the reversible photo bleaching.

For the measurements presented here, data stacks consisting of several thousand images were recorded with an integration time of the camera of 50 to 150 ms. After conversion of the count numbers into photon numbers, a differential image stack was calculated by subtracting the succeeding from the preceding frame. This had to be done to filter out the signals of the single molecules due to background noise and bleaching gradients of biological samples. After segmentation of the raw data, a 2D Gaussian was fitted to the single molecule signals to determine the positions of the detected molecules [4,15,16,24]. Using this information, a localization image was rendered by blurring the position of each detected molecule with a Gaussian corresponding to the individual localization accuracy.