The cold embrace of death: using time-lapse imaging to understand what kills leaves when they freeze

Abstract

This article comments on:

Johnson KM, Scherer M, Gerber D, Style RW, Dufresne ER, Brodersen CR. 2025. Ice and air: visualisation of freezing spread and freeze–thaw embolism in young Liriodendron tulipifera leaves. Journal of Experimental Botany 76, 5573–5587. https://doi.org/10.1093/jxb/eraf254

This article comments on:

Johnson KM, Scherer M, Gerber D, Style RW, Dufresne ER, Brodersen CR. 2025. Ice and air: visualisation of freezing spread and freeze–thaw embolism in young Liriodendron tulipifera leaves. Journal of Experimental Botany 76, 5573–5587. https://doi.org/10.1093/jxb/eraf254

Plants are mostly ectothermic. To cope with both intermittent and long-term freezing stress, they have developed a variety of avoidance and tolerance strategies. Johnson et al. (2025) adapt an imaging technique typically associated with measuring drought tolerance to visualize freezing in the young leaves of Liriodendron tulipifera L. (Magnoliaceae). Theses imaging data are paired with anatomical measurements to better understand how leaf structure influences ice spread within leaves and contributes to freeze–thaw embolism. The findings by Johnson et al. (2025) represent a significant step toward elucidating how deciduous plants protect themselves from unseasonal freezing events and offer the first visualization of freeze–thaw embolism in leaves.

Freezing is a major stressor for plants and plays a crucial role in determining plant distribution and leaf habit (Shreve, 1914). Deciduous plants, such as the north-American Liriodendron tulipifera (the state tree of Indiana, Kentucky, and Tennessee), shed their leaves in winter because deciduous leaves are not typically built to tolerate the freezing temperatures during what is usually a photosynthetically unproductive time of year (Wolfe and Upchurch, 1986; Givnish, 2002). Freezing not only limits the natural ranges of wild plants but is also a major consideration in the human cultivation and distribution of plants for agriculture and landscaping. Although much work has gone into improving cold and frost tolerance in crop species to expand the regions where they can be grown (Jan et al., 2024), most species remain relatively confined to particular temperature regimes. As the climate continues to warm, it may seem that frost stress would become a diminishing concern. However, the combination of earlier spring leaf-out and more frequent late-season frosts has created a ‘perfect storm’, where young, tender leaves are increasingly exposed to severe frost events (Lamichhane, 2021). These frosts can be devastating for both natural forests and agricultural systems, leading to substantial losses in productivity and yield. In some cases, trees must completely regenerate canopies within a growing season, significantly impacting annual carbon gain. From an agricultural perspective, freezing events cause greater economic losses in the USA than any other weather phenomenon, with short durations of cold resulting in hundreds of millions of dollars in lost crops and untold downstream effects on local economies (Snyder and de Melo-Abreu, 2006). Despite the growing importance of this issue, research on ice nucleation and propagation in plant tissues has been limited in recent years (Wisniewski et al., 2009). This makes the work presented in Johnson et al. (2025) all the more important as, in combination with other recent work (Kane and McAdam, 2024), to adapt field-deployable time-lapse cameras to visualize freezing dynamics in leaves at fine spatial scales provides insights into how photosynthetic tissues resist or succumb to ice-induced damage.

Time-lapse imaging has long been a powerful scientific tool. The earliest use of time-lapse photography is often attributed to Eadweard Muybridge in 1872, who employed it to study horse locomotion, providing increased data for improving racehorse training (Leslie, 2001). Since then, various forms of time-lapse photography have been used by scientists to capture processes that are either too slow or too fast to be observed by the naked eye. In plant biology, time-lapse imaging has been used to document phenomena across a wide range of scales from diurnal plant movements (Rehman et al., 2020), to the dynamics of stomatal opening and closing (Kappen et al., 1987), and even to satellite-based tracking of leaf phenology (MacBean et al., 2015). Within the field of plant hydraulics, time-lapse optical methods have become increasingly common for capturing xylem embolism formation in stems (Brodribb et al., 2017), leaves (Brodribb et al., 2016), and roots (Rodriguez-Dominguez et al., 2018), allowing researchers to quantify the ability of species to withstand drought. The advent of relatively low-cost, field-deployable, Raspberry Pi-powered ‘Cavicams’ (https://cavicams.com/) has further expanded the accessibility of this technique. These cameras have recently been used to monitor the shrinking and swelling of petioles as a proxy for plant water potential (Bourbia and Brodribb, 2023). More recently, the studies by Kane and McAdam (2024) and Johnson et al. (2025) have adapted Cavicams to observe freezing dynamics and the spread of ice across leaves, opening up new opportunities for affordable, in-field observation of plant responses to freezing stress.

Johnson et al. (2025) exposed saplings of L. tulipifera with juvenile leaves to controlled freezing events designed to mimic a natural late spring frost. These leaves were monitored using optical vulnerability time-lapse cameras, which are typically used to visualize xylem embolism during drought stress (Brodribb et al., 2016, 2017; Cardoso et al., 2022). Instead of drought-induced embolism, Johnson et al. (2025) captured the dynamic spread of ice through the leaf mesophyll, beginning in the ground tissue surrounding the midrib and secondary veins, and eventually spreading throughout the rest of the leaf. Notably, the major veins were the last to freeze—similar to observations in Lonicera × purpusii (Kane and McAdam, 2024). Ice spread in the mesophyll was often patchy, leaving small regions of unfrozen tissue surrounded by ice. Using chlorophyll fluorescence imaging, the authors found that tissue damage indicated by impaired photosynthesis occurred primarily in regions that froze, while unfrozen patches maintained high fluorescence and appeared to survive with minimal impairment, suggesting that it is the formation of ice that is causing mesophyll dysfunction. The paired anatomical measurements conducted by Johnson et al. (2025) show that bundle sheath extensions (BSEs) around veins may help prevent the lateral spread of ice through the leaf lamina, a mechanism previously observed only in Cinnamomum camphora (Hacker and Neuner, 2007). In L. tulipifera, BSEs were present around tertiary veins but not around primary or secondary veins, which may explain the rapid initial spread of ice near the midrib and secondary veins and the subsequent slowing as ice moved into the lamina. In addition, Johnson et al. (2025) present the first in situ observations of freeze–thaw-induced embolism captured using the optical method. These embolism events occurred over the course of minutes after thawing, whereas comparable embolism levels during drought typically require hours. Previous work has shown a strong relationship between vessel diameter and susceptibility to freeze–thaw embolism, with wider vessels being more prone to embolize (Pittermann and Sperry, 2003). Consistent with this, Johnson et al. (2025) show that no embolism was observed in vessels <15 μm in diameter, indicating that all freeze–thaw-induced embolism was restricted to the midrib and secondary veins.

The use of pixel brightness, whether from RGB or gray-scale images, has proven to be a robust proxy for tracking ice spread in leaf tissues (Kane and McAdam, 2024; Johnson et al., 2025). Johnson et al. (2025) suggest that changes in pixel brightness are probably driven by the formation of ice in intercellular spaces, accompanied by the movement of water out of cells and into those spaces to join the growing ice crystals as temperatures drop. This results in a combination of freezing and drought stress, which aligns with current understanding of how plants avoid mesophyll cell death during freezing events (Steponkus and Webb, 1992). At the cellular level, utilizing high-resolution confocal microscopy and reflected light, Johnson et al. (2025) observed darkening of cells during cooling, indicating dehydration and supporting the idea that water was indeed exiting the cells, which is consistent with the early observations made by Stuckey and Curtis (1938) in seeds. Time-lapse imaging may also offer valuable insights into post-thaw leaf damage. Beyond visualizing embolism, Johnson et al. (2025) noted a reddening of leaves upon thawing. Similar reddening or browning has been observed in other species following freezing events, sometimes attributed to anthocyanin production (Ball et al., 2002). However, due to the rapid onset of color change observed by Johnson et al. (2025), it is more likely to be the result of oxidation of cellular components caused by mesophyll damage. This rapid discoloration could serve as a useful non-invasive indicator for assessing the extent of freezing damage to leaves (Fig. 1).

Fig. 1.

Decreases in pixel greenness reveals leaf damage upon thawing. (A–D) Images of a leaf of Lonicera × purpusii at different time points (A) pre-freeze, (B) frozen, (C) thawing, and (D) thawed, while being rapidly frozen and thawed at a rate of ∼2°C min^–1 and ∼0.6°C min^–1, respectively in a –10 °C freezer. (A) Scale bar showing 1.25 mm. Images were captured every 40 s with a Raspberry Pi-powered Cavivam. The images are subdivided into upper and lower halves that correspond to the blue and red bars to the left of (A). (E) Graph showing the average pixel greenness of each image in the sequence extracted using ImageJ according to Kane and McAdam (2024). The blue line corresponds to the upper half of the images, while the red line corresponds to the average pixel greens of the lower half of the image sequence. The dashed line shows leaf temperature as measured by a fine wire thermocouple once a minute. The vertical black bars correspond to the time points of the representative images (A–D), and the * shows the observed exotherm.

Freeze–thaw embolism is a significant concern for plants growing in temperate and alpine regions, yet it has received far less attention than drought-induced embolism. Johnson et al. (2025) are the first to visualize freeze–thaw embolism in leaves—an important step toward answering key questions about the frequency and severity of embolism during freezing events. It has been proposed that there is a trade-off between frost and drought tolerance (Larter et al., 2024), and the methods presented by Johnson et al. (2025), when combined with conventional optical vulnerability techniques, now make it much easier to test this hypothesis. This work also paves the way for future research into how freezing events affect water conductance in temperate trees before and after winter. Notably, Sperry et al. (1994) found that in dicotyledonous species, > 90% of stem embolism occurred during winter, driven primarily by freeze–thaw events. With the advent of affordable and deployable imaging systems, researchers can now observe freeze–thaw embolism dynamics across a broader range of species and environmental conditions, opening up new avenues for understanding plant hydraulic function in cold climates.

Johnson et al. (2025) present a clear picture of the freezing dynamics in young leaves and find that the apparent driver of leaf lethality during freezing is mesophyll disruption and, maybe to a much lesser extent, xylem embolism, with embolism observed in far fewer repetitions than damage to photosynthetic tissues which were seen in all freezing trials. This invites the question as to how this dynamic may shift as leaves mature and acclimate, which can be investigated with the new uses of technology presented here. This, along with other recent work using a frost-tolerant plant (L.×purpusii) (Kane and McAdam, 2024) in which freeze–thaw embolism was not observed, raises the possibility that freeze–thaw embolism might only be happening in plants and tissues that are not designed to survive the winter.