FIGURE SUMMARY
Title

Purine Molecular Interactions Determine Anisotropic Shape of Zebrafish Biogenic Crystals

Authors
Rothkegel, J., Kaufmann, S., Wilsch-Bräuninger, M., Lopes, C., Mateus, R.
Source
Full text @ Small Methods

Crystal reflection onset and morphology in larval zebrafish. a–c) Confocal microscopy maximum intensity projection images of larvae eyes of iridophore reporter line (TDL358:GFP, green) and reflection (white), at 48 hpf (a), 56 hpf (b) and 72 hpf (c), with zoom in on crystals (a’‐c’). Cartoon indicates crystallographic axes in 2D with respective crystal facets. Scale bars: a‐c 50 µm, a’‐c’ 10 µm. d,e) Iridophores’ area in the eye (d) increases linearly between 30 and 96 hpf (d; Line, linear fit with goodness of fit, R2), while iridophore reflection (e) increases exponentially within the same time interval, in the same cells (e, Line, exponential fit with goodness of fit, R2). Mean ±SEM, black lines. f–i) Morphometric measurements from 2D segmented crystals obtained from reflection imaging of wild‐type zebrafish eyes at three different time points. Area of (100) crystal facet (f), Aspect Ratio of b‐axis/c‐axis lengths (g), b‐axis length (h), and c‐axis length (i). Asterisks indicate p‐values of unpaired, two‐tailed, non‐parametric t‐tests between groups at each time point: p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Error bars, standard deviation of the mean. N, number of crystals. 48 hpf, N = 197; 56 hpf, N = 227; 72 hpf, N = 107. j) CryoFIB‐SEM slice of wildtype iridophore at 72 hpf. Guanine crystals appear dark due to their high density (black arrow) in comparison to the rest of the cell. The FIB‐milling and imaging direction are parallel to the crystals’ a‐axis (inset). Dash, eye surface. Green line, manual annotation of the cell membrane. Magenta line, manual annotation of the cell nucleus. Scale bar, 10 µm. k) 3D crystal reconstruction from (j). Segmented crystals are colored relative to volume, with darker colors corresponding to lower volumes. Cell outline (green) and nucleus (magenta) correspond to the ones depicted in (j). See also Movie S1 (Supporting Information). l) 3D segmentation mask (left) and respective volumetric rendering (right) of a single wildtype crystal (from k), with crystallographic axes and respective facets annotated. Scale bar: 1 µm. m–p) Measurements from 3D segmented crystals obtained from cryoFIB‐SEM imaging; N = 139 crystals. Volume (m), b‐axis length (n), c‐axis length (o), and a‐axis length (p).

Pnp4a−/− crystals display distinct square‐like macromorphology due to an underdeveloped (001) crystal facet. a) Eye iridophores labeled in double transgenic TDL358:GFP (green) and pnp4a:PALM‐mCherry (magenta), with zoom highlighting co‐localization of expression and crystals’ reflection (white), at 72 hpf (a′‐a″). Scale bars: 50 µm (a), 10 µm (a’‐a″). b‐b’ Maximum intensity projection of confocal reflection images of larval eyes in wildtype (b) and Pnp4a−/‐ (b’) at 72 hpf. Insets highlight respective crystal morphologies. Scale bars: 100 µm, 5 µm (inset). c–f) Comparisons of 2D morphometric crystal measurements between wildtype (blue) and Pnp4a−/− mutants (red) at 56 and 72 hpf. Area of (100) crystal facet (c), Aspect ratio between lengths of b‐ and c‐axis (d), b‐axis length (e), and c‐axis length (f). N, number of crystals. 56 hpf, WT N = 227, Pnp4a−/−N = 41; 72 2.7kb‐hpf, WT N = 107, Pnp4a−/−N = 315. g) CryoFIB‐SEM slice of Pnp4a−/− iridophore at 72 hpf. Black arrow, guanine crystal. The FIB‐milling and imaging direction are parallel to the crystals’ a‐axis. Dash, eye surface. Green line, manual annotation of the cell membrane. Magenta line, manual annotation of the cell nucleus. Scale bar, 10 µm. h) 3D segmentation mask (left) and respective volumetric rendering (right) of a single Pnp4a−/− crystal, with crystallographic axes and respective facets annotated. h’ While crystals in Pnp4a−/‐ mutants express square‐like habits, the extent of their underdeveloped (001) facets presents a range of variability. Scale bars: 1 µm. i–m) Comparison of 3D morphometric measurements between wildtype (blue) (N = 139) and Pnp4a−/− crystals (red) (N = 10). Volume (i), Aspect ratio of b‐axis/c‐axis lengths (j), b‐axis length (k), c‐axis length (l), and a‐axis length (m). Asterisks indicate p‐values of unpaired, two‐tailed, non‐parametric t‐tests between groups: p < 0.05, ∗∗p < 0.01, ∗∗∗∗p < 0.0001. Error bars, standard deviation of the mean.

Zebrafish crystal macromorphology depends on H‐bond interaction strength between iridosomal purines. a) Chemical structure of guanine (top) and hypoxanthine (bottom). The two molecules only differ by a ‐NH2 group (red box). b) b‐c plane of the β‐guanine crystal structure. The amino group (red box) forms two hydrogen bonds (cyan) with neighboring guanine molecules. c) b‐c plane of β‐guanine crystal structure without the identified ‐NH2 hydrogen bond interactions (red box in b) corresponds to zero interaction strength in (e). d) b‐c plane of β‐guanine crystal structure with the central guanine molecule being exchanged by hypoxanthine (shaded red molecule). This hypothetical conformation does not have the two hydrogen bonds identified in the pure guanine lattice (red box in b). Like in (c), this corresponds to a scenario where hydrogen bond interactions would have zero cooperation. e) From left to right, simulated crystal morphologies from (c) in decreasing ‐NH2 interaction strength order (range 0–5). Note the striking similarity of simulated crystals with zero interaction strength with the obtained in vivo Pnp4a−/− crystals. f) From left to right, 3D rendered cryoFIB‐SEM crystals aligned with the closest resembling simulated crystal appearance. Scale bar: 1 µm. See also Figure S1e (Supporting Information).

In vivo Pnp5a overexpression leads to smaller (100) crystal facets. a,b) Maximum intensity projections of larval eyes highlighting iridophore reflection upon heat‐shock overactivation of Pnp5a transgene (hsp70:Pnp5a‐GFP, green) (b‐b’) or in heat‐shocked sibling controls (a‐a’), at 72 hpf. In both cases, iridophores are labeled by the 2.7kb‐tfec:mKate2 (magenta) reporter line. Insets (a″‐b″) highlight respective crystal morphologies. c,d) Maximum intensity projections of larval eyes highlighting iridophore reflection in Pnp5a homozygous mutants (Pnp5a−/−, d‐d’) or in wildtype controls (c‐c’), at 72 hpf. In both cases, iridophores are labeled by the TDL358:GFP (green) reporter line. Insets (c″‐d″) highlight respective crystal morphologies. Scale bars: 50 µm (a–d), 5 µm (a″‐d″). e,f) Comparison of the average iridophore reflection per 50 µm in different WT, Pnp5a overexpression (e) or mutant conditions (f), at 72 hpf. au, arbitrary units. Mean ±SEM are shown in black. Heat‐shock sibling control, N = 5; Pnp5a overexpression, N = 10; WT, N = 10; Pnp5a−/−, N = 15. Asterisks indicate p‐values of two‐tailed, non‐parametric Welch t‐tests between groups: p < 0.05, ∗∗p < 0.01.

Acknowledgments
This image is the copyrighted work of the attributed author or publisher, and ZFIN has permission only to display this image to its users. Additional permissions should be obtained from the applicable author or publisher of the image. Full text @ Small Methods