Laboratory cultures of the C. watsonii strain (WH8501) were grown in SO and SN media [12; with and without combined nitrogen] in a 12:12 light dark cycle at 28 °C. Cultures of the T. erythraeum strain (IMS101) were grown on YBC11 media [13] in a 12:12 light dark cycle at 28 °C. The cultures were grown in either iron deplete or iron replete conditions. The iron deplete cultures had iron removed from the trace metal mix in order to induce accumulation of the iron stress protein IdiA.

Cells of C. watsonii (2 mL) were immunolabeled using a modified method derived from [14,15]. The single cell cyanobacteria were collected by filtration on a 1.2 μm filter and placed into ice cold 95% ethanol at −20 °C overnight. The C. watsonii cells were always collected in the middle of the dark period to ensure maximum expression of the nitrogenase protein. Cells were centrifuged and fixed in 1% paraformaldehyde (pH 6.8, buffered with 100 mM PBS) at room temperature for 15 min, and incubated in lysozyme (10 mg/mL in 100 mM Tris-HCL, 20 mM EDTA, pH 8.0, Sigma) for 1 h at 37 °C, followed by achromopeptidase (60 U/mL in 50 mM Tris-HCL, pH 7.0, Sigma) for 30 min at 37 °C to permealize the cells. The permealization method was tested on centrifuged cell suspensions with the Bac/Light Live Dead stain (Invitrogen). Positive results (red stained cells) were obtained showing the integrity of the membrane had been perforated. The cells were blocked in 3% bovine serum albumin (BSA, Sigma) for 1 h at room temperature. After blocking, the cells were incubated in polyclonal primary antibody solution rabbit IgG (anti-IdiA or anti-nitrogenase) at a titer of 1:100 for 1 h at room temperature. The cells were then incubated with secondary Qdot 525 or Qdot 605 goat anti-rabbit IgG (Invitrogen) 1:100 solution overnight at 4 °C. After each antibody incubation, the cells were rinsed once with 0.05% Tween in PBS followed by a rinse in PBS. Cells were incubated in Qdot 525 or Qdot 605 goat anti-rabbit IgG 1:100 solution without primary antisera as a control for non-specific binding. Each step unless indicated in the procedure, had three rinses with PBS. Cells grown in SN media were used as a control. Filaments of T. erythraeum were collected by filtration (3 mL) on a 2 μm filter and suspended in ice cold 95% ethanol at −20 °C overnight. The filaments were collected on a 2 μm filter, placed in a plastic Petri dish and the filter encircled with a PAP pen (Energy Beam Sciences, Inc.). The filaments were permeabilized with lysozyme (10 mg/mL) for 3 min at 37 °C followed by 0.5% dimethyl sulphoxide (DMSO, Sigma) for 15 min at 4 °C. The Trichodesmium filaments were blocked in 3% BSA for 1 h prior to incubation in 1:100 titer primary anti-nitrogenase or anti-IdiA sera for 4 h at room temperature. Finally, the filaments were incubated in secondary 525 Qdot goat anti-rabbit IgG sera overnight at 4 °C. The C. watsonii cells were photographed immediately to document autofluorescence and then again after exposure to blue light (450–480 nm) for 60 s using an Olympus BX51 fluorescent microscope equipped with a DP71 12 bit color digital camera and software. The T. erythraeum filaments however, required an additional pretreatment of 4–6 hour exposure to short wave UV to reduce autofluorescence before blue light exposure for 15 min. All preparations were viewed using an Olympus U-MWB2 epifluorescent microscope with a wide blue filter cube with excitation at 450–480 nm, dichroic mirror 500 nm and longpass barrier filter at 515 nm.

Iron stressed C. watsonii cells probed with the IdiA antibody and labeled with Qdot 605 secondary antibody probe that were exposed to blue light for 60 s displayed a bright fluorescent signal from the photostable nanocrystal conjugate (Figure 1B). T. erythraeum trichomes grown in iron replete media did not show any labeling of IdiA protein (Figure 2A,B) following 4 h exposure to short wave UV and subsequent 15 min exposure to blue light. T. erythraeum trichomes grown on iron deplete media showed IdiA labeling along the trichomes in many cells (Figure 2C,D) following the same pretreatment as in iron replete cultures. Iron stressed trichomes were associated with cellular debris or exudates that incorporated some of the Qdot antibody conjugate in the extracellular matrix (Figure 2C,D). Omitting the primary antibody showed no labeling in T. erythraeum cells (data not shown). This fluorescent immunolabeling technique would be impossible with conventional organic fluorophores as the natural autofluorescence (Figure 1A, Figure 2A,C) would mask the antibody signal. Also, organic fluorophores rapidly fade under blue light exposure and especially prolonged exposure to short wave UV. Thus, the photostability of Qdot antibody probes allows for unprecedented fluorescent immunodetection in highly autofluorescent cells. The interference of the natural blue fluorescence found in unicellular cyanobacteria [1,5] was avoided by using the Olympus WB2 wide blue filter cube with emission greater than 515 nm. Unlike conventional organic fluorophores, Qdots have a wide absorption spectrum, a narrow, tunable emission spectrum, and are photo-chemically stable. The unique optical properties of Qdots make them ideal probes for highly autofluorescent cyanobacterial cells.

Cells of C. watsonii did not show any non-specific binding of the Qdot conjugate when the primary antiserum was omitted (Figure 3A,B). Cells grown in iron replete media only showed trace labeling of IdiA (Figure 3C,D) while cells grown in iron deplete media showed the most pronounced labeling (Figure 3E,F). This whole-cell approach can detect iron stress, expressed as IdiA, in single cells and filaments of cyanobacteria providing a useful and powerful probe for physiology studies of open ocean cyanobacteria.

Up until now, fluorescent immunolabeling of nitrogenase in cyanobacteria has been problematic due to autofluorescence [1–3,5]. Additionally, the use of fluorescent probes in unicellular cyanobacteria are frequently masked by a natural blue fluorescent pigmentation [1,5]. Due to its photostability, broad absorption spectrum and tunable emission pattern, Qdot antibody probes can overcome these issues. Nitrogenase was immunolabeled with Qdot conjugate 605 in C. watsonii (Figure 4) and Qdot conjugate 525 in T. erythraeum (Figure 5A,B). The C. watsonii cells were exposed to blue light for 60 s to reduce autofluorescence. Trichodesmium filaments however, took longer to photobleach the natural pigments and required additional hours of short wave UV exposure to reduce autofluorescence. In most cases, extended exposure to UV light would photobleach conventional organic fluorophores, but the Qdot antibody conjugate yielded a bright luminescent signal. Cells of C. watsonii grown in SN medium did not show significant labeling (data not shown). All cyanobacterial cells incubated in Qdot secondary antibody solution without primary antibody to nitrogenase did not show any labeling with Qdot conjugates (data not shown). This is the first report of whole-cell immunolocalization of nitrogenase and IdiA using Qdot conjugates in both filamentous and single cell cyanobacteria.

Two different labeling patterns were observed in T. erythraeum cells. Figure 5A shows a pattern that is similar to recent work using an improved immunocytochemical technique [2,3] and immungold labeling [20]. This pattern showed labeling of nitrogenase that was less intense (relative to Figure 5B), more diffuse and appeared in a greater number of cells along the trichome (Figure 5A). The pattern in Figure 5B showed intense labeling of nitrogenase in a subset of cells along the trichome. This pattern is similar to previous work using Alexa-350 fluorochromes [4,16,17] and to studies using immunogold labeling [18,19]. The difference in labeling pattern might be attributed to age or physiology as suggested by Ohki and Taniuchi [3] since the cells in Figure 5B were 20 days older than the cells in Figure 5A. The trichomes in Figure 5B had formed bundles in culture while the trichomes in Figure 5A were free trichomes. Bundle formation may be a precursor to consolidation of the enzyme in a subset of cells. The use of Qdot antibody conjugates in this study clearly demonstrates the usefulness of these fluorophores in detecting low abundant cellular antigens as well as intense labeling of filaments in highly autofluorescent cyanobacteria such as T. erythraeum.