Immersion microscopy at cryogenic temperature

Here, in situ cryofixed C. elegans were imaged in confocal microscopy at about -145˚C, using a high-NA immersion objective (63x/1.15 NA) that can work below the glass transition temperature of water (-137˚C). The results of these experiments show that the transfer process faithfully preserves the shape of the sample and enables the acquisition of cryofluorescence images with significantly improved contrast and resolution. Cryofixed samples were provided by Marie Fuest, PhD and Rodrigo Galilea Kleinsteuber. Data in cryo-immersion were acquired in collaboration with Margherita Bassu, PhD.

2.1. C. elegans shape is preserved after sample transfer

Removing the sample from the microfluidic cryofixation system poses the risk of recrystallization or mechanical damage. Here, channel features (walls, traps) were used as fiducial markers to assess whether the sample transfer damaged the sample quality. The channel features from the cryofixation frame and in confocal immersion cryo-microscopy were superimposed to compare the shape of the nematode before and after the transfer step (Figure 35).

Figure 35 Overlay of the nematodes before and after transfer (cryofixation frame and immersion cryo-microscopy). The superposition of the channel features showed good alignment of the nematode body, indication of good preservation of the sample throughout the transfer steps. Cryofixed samples were provided by Marie Fuest, PhD and Rodrigo Galilea Kleinsteuber. Data in cryo-immersion was acquired in collaboration with Margherita Bassu, PhD. Scale bar measures 30 µm.

From the absence of changes in the shape of the nematodes, the preservation quality of the samples was evaluated to be good with no detectable damage.

Figure 36 C. elegans cryofixed in the microfluidic channel and imaged in confocal microscopy at ~-145˚C using a cryoimmersion objective (63x/1.15 NA). Green, GCaMP; red, gut granules. Cryofixed samples were provided by Marie Fuest, PhD and Rodrigo Galilea Kleinsteuber. Data in cryo-immersion was acquired in collaboration with Margherita Bassu, PhD. Scale bar measures 30 µm.

As observed in live imaging and low-NA cryo-microscopy, two fluorescent objects are visible in the sample: the muscle fibers (green channel) and the autofluorescent gut granules (red channel).

Acquiring multiple FOVs, it was possible to reconstruct the nematode body. From the relative brightness of the image, it was also possible to see which muscles were contracting in the moment of freezing. Indeed, the increase in calcium concentration that triggers the muscle contraction also makes the GCaMP molecules more fluorescent [Tian et al., 2009] (Figure 37, green arrows).

Figure 37 Reconstruction of nematode body stitching three FOVs acquired in fluorescence immersion cryo-microscopy. Fluorescence signal comes from GCaMP molecules expressed in the muscle fibers and from autofluorescent gut granules. Brighter muscle fibers (arrows) are associated with more contraction at the moment of fixation. Cryofixed sample was provided by Marie Fuest, PhD and Rodrigo Galilea Kleinsteuber. Data in cryo-immersion was acquired in collaboration with Margherita Bassu, PhD. Scale bar measures 30 µm.

2.2. Image quality in confocal cryoimmersion microscopy

As previously introduced, the low numerical aperture of the optics of the cryofixation device limits the achievable image quality. Here, a significant improvement in image quality was attained by transferring the sample to a confocal cryoimmersion microscope.

Although kept under liquid nitrogen until moved to a dry cryogenic environment for imaging, samples showed frost build up, in particular near the channel outline (Figure 38). The image quality observed with the air objective deteriorated due to frost formation. However, this was alleviated once the sample was observed in immersion cryo-microscopy. The lack of detection of frost in immersion cryo-microscopy was probably due to the matching of the refractive index of the immersion fluid with the refractive index of the ice.

Figure 38 Microfluidic chip containing cryofixed C. elegans imaged in a dry cryogenic environment (10x/0.2 NA). In fluorescence, it is possible to see the outline of the nematode in the middle of the microfluidic channel.

Fluorescence signal comes from GCaMP and autofluorescent gut granules. The presence of frost, that prevents good imaging when an air objective is used, is alleviated likely by the matching of water refractive index in immersion cryo-microscopy. Scale bar measures 300 µm.

The gray values of a line traced across the nematode body were compared between the cryofixation frame and the immersion cryo-confocal image. Plotting the normalized values against the length of the line, the change in contrast was assessed. The normalized intensity along the traced line (𝑰̅(𝒙)) was computed using the following equation:

𝑰̅(𝒙) =𝑰(𝒙) − 𝑰_𝒊 𝑰_𝒃𝒈

( 8 )

Where 𝐼_𝑖 is the intensity measured in the middle of the nematode body and 𝐼_𝑏𝑔 is the intensity of the background as schematized in Figure 39.

The contrast attained in confocal cryoimmersion microscopy was assessed to be about 20 times higher than in widefield microscopy (Figure 39).

Figure 39 Measured intensity across nematode body in widefield and in confocal cryomicroscopy. In confocal cryomicroscopy, the contrast measured across the nematode body was about 20 times higher than in widefield microscopy. Data in cryo-immersion was acquired in collaboration with Margherita Bassu, PhD.

The two peaks in Figure 39 appear to have different relative intensities and smaller full width at half maximum (FWHM) in cryo-confocal microscopy compared to widefield microscopy.

However, these measurements can be biased by the uncertainty along the z-axes of the sample.

In fact, due to the round shape of the nematode and its thickness, different image planes will have different relative intensities on the same region of interest.

For instance, in the right panel of Figure 35 three consequences of this aspect can be remarked. First, the muscle groups of the head of the nematode, that appear strongly fluorescent in widefield, are not visible in confocal. Second, only few granules are visible in cryo-confocal whereas there seem to be many more in widefield. Third, the muscles seem to run in the inner part of the nematode body rather than on the edge, although perfectly following its contour. All three of these observations can be explained by the position of the image plane in the upper part of the nematode body rather than the middle plane.