Liu et al.

Stress Biology (2025) 5:39 Stress Biology

https://doi.org/10.1007/s44154-025-00239-4

ORIGINAL PAPER Open Access

The mRNA‑binding protein HLN1 enhances

drought stress tolerance by stabilizing the GAD2
mRNA in Arabidopsis
Chuangfeng Liu1, Yang Wang2, Jialin Peng1, Zhengyu Shao1, Yajie Liu1, Zhiqing Zhang1, Xiaoyu Mo1, Yilin Yang1,
Tao Qin3, Yiji Xia1,4 and Liming Xiong1,4*   

Abstract
Drought is a common environmental condition that significantly impairs plant growth. In response to drought, plants
close their stomata to minimize transpiration and meanwhile activate many stress-responsive genes to mitigate dam-
age. These stress-related mRNA transcripts require the assistance of RNA-binding proteins throughout their metabolic
process, culminating in protein synthesis in the cytoplasm. In this study, we identified HLN1 (Hyaluronan 1), an RNA-
binding protein with similarity to the animal hyaluronan-binding protein 4 / serpin mRNA binding protein 1 (HABP4/
SERBP1), as crucial for plant drought tolerance. The hln1 loss-of-function mutant exhibited higher transpiration rates
due to impaired stomatal closure, making it highly susceptible to drought. Drought stress increased HLN1 expression,
and the protein underwent liquid–liquid phase separation (LLPS) to form mRNA-ribonucleoprotein (mRNP) conden-
sates in the cytoplasm under osmotic stress. We identified GAD2 as a potential mRNA target of HLN1. GAD2 encodes
the predominant glutamate decarboxylase synthesizing γ‐aminobutyric acid (GABA), a non-proteinogenic amino
acid that modulates stomatal movement. RIP-qPCR and EMSA showed that HLN1 binds GAD2 mRNA, which promotes
HLN1 condensate formation. In hln1 mutants, GAD2 transcripts were less stable, reducing steady-state mRNA levels.
As a result, hln1 accumulated less GABA and exhibited impaired stomatal closure under drought. Conversely, HLN1
overexpression stabilized GAD2 mRNA, increased GABA levels, and enhanced drought tolerance in transgenic plants.
GAD2 overexpression in hln1 mutants also rescued the drought-sensitive phenotypes. Overall, our study reveals
a mechanism whereby HLN1 stabilizes GAD2 mRNA to enhance GABA production and drought tolerance. These find-
ings provide novel strategies for engineering drought-resistant crops.
Keywords Drought, mRNA-binding protein, Condensate, mRNA stability, HLN1, GABA

Introduction
Drought stress significantly reduces plant growth and
crop productivity. To cope with drought, plants have
Handling Editor: Huazhong Shi.
evolved diverse adaptive responses, such as develop-
*Correspondence: ing deep root systems for better water foraging, closing
Liming Xiong
lxiong@hkbu.edu.hk stomata to reduce transpiration, and activating thou-
1
Department of Biology, Hong Kong Baptist University, Kowloon Tong, sands of stress-responsive genes whose products mitigate
Hong Kong, China stress damage (Zhu 2016; Gong et al. 2020; Sato et al.
2
College of Forestry, Northwest A&F University, Yangling 712100, China
3
College of Grassland Agriculture, Northwest A&F University, 2024). These discrete yet intertwined responses collec-
Yangling 712100, China tively enhance plant drought tolerance. Among these
4
State Key Laboratory of Agrobiotechnology, The Chinese University responses, stomatal closure and stress gene expression
of Hong Kong, Sha Tin, Hong Kong, China
occur rapidly upon osmotic stress induced by drought

© The Author(s) 2025. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://creativecommons.org/licenses/by/4.0/.
Liu et al. Stress Biology (2025) 5:39 Page 2 of 17

and are tightly regulated. In particular, stress gene activa- GAD2 overexpression or exogenous GABA supplementa-
tion requires coordinated regulation across transcription, tion rescued the impaired stomatal closure and drought-
processing, and translation, with RNA-binding proteins sensitive phenotypes of the hln1 mutant. In summary,
(RBPs) playing critical roles at each stage. Although thou- our study reveals that HLN1 enhances drought tolerance
sands of RBPs have been annotated in plant genomes by stabilizing GAD2 mRNA and promoting GABA-medi-
(Ambrosone et al. 2015; Marondedze et al. 2019; Sajeev ated stomatal regulation under drought stress.
et al. 2022; Wang et al. 2024), only a limited number have
been directly linked to stress responses, and their mecha- Results
nisms remain poorly understood. Given the importance HLN1 is required for drought stress tolerance
of RBPs in gene regulation, elucidating their functions in Upon drought stress, plants reduce stomatal conductance
stress responses is vitally important. to minimize water loss, leading to elevated leaf tempera-
The RNA-binding capacity of RBPs is determined by tures. Using thermal imaging, we isolated an Arabidop-
their sequence and structural features. Beyond canoni- sis mutant, lot1 (lower temperature 1), which exhibits
cal RNA-binding domains, many RBPs possess intrinsi- higher transpiration rates and greater drought suscep-
cally disordered regions (IDRs) that have the propensity tibility (Qin et al. 2019). LOT1 encodes a protein poten-
to undergo liquid–liquid phase separation (LLPS) driven tially involved in the posttranslational modifications of
by multivalent interactions, particularly under cellular stress-signaling proteins, regulating their localization
crowding, to form membraneless condensates. These and stability (Qin et al. 2019). To elucidate the modes of
condensates comprise specific RBPs, associated mRNAs action of LOT1, we performed yeast two-hybrid screens
and other proteins, yielding distinct granules such as (Qin et al. 2019) and conducted co-immunoprecipitation
stress granules (SG) and processing bodies (P-bodies), assays with tagged LOT1 transgenic plants. These experi-
which compartmentalize cellular activities. By concen- ments identified the RNA-binding protein HLN1 (Hyalu-
trating components for transcript processing, storage, ronan 1, AT5G47210) as a potential component involved
modification, translation, or degradation, these dynamic in LOT1-mediated drought responses. We thus investi-
structures facilitate gene regulation (Hentze et al. 2018; gated further the potential functions of HLN1 in modu-
Kosmacz et al. 2019; Emenecker et al. 2020; Wiedner and lating drought stress responses.
Giudice 2021; Hirose et al. 2023; Wadsworth et al. 2024). We obtained an hln1 knockout mutant and two HLN1
While current knowledge about biomolecular conden- complementation lines (Comp #1 and #2, i.e., pHLN1-
sates is mostly derived from non-plant systems, the prev- HLN1/hln1 #1 and #2, expressing HLN1 genomic DNA
alence of stress-responsive genes in plants suggests that under its native promoter in the hln1 mutant back-
RBP-driven condensates likely play critical roles in stress ground) to assess drought responses. Three-week-old
signaling and adaptation in plants. soil-grown plants were subjected to 20 days of water
In this study, we identified an RNA binding protein deprivation followed by re-watering. Two days later, the
Hyaluronan 1 (HLN1) as an important positive regula- plants were photographed (Fig. 1A). While wild type
tor of plant drought tolerance. The hln1 mutant exhibited (Col-0) plants survived, the hln1 mutant did not. Notably,
elevated transpiration and heightened drought-sensi- both complementation lines rescued the hln1 mutant’s
tivity. HLN1 is an RNA-binding protein with similarity drought sensitivity, confirming that the phenotype
to animal hyaluronan-binding protein 4 (HABP4) and resulted from HLN1 loss (Fig. 1A).
serpin mRNA binding protein 1 (SERBP1), and it har- Stomatal conductance of these plants was monitored
bors multiple predicted IDRs. We demonstrated that using a porometer. Significant differences in stomatal
HLN1 undergoes LLPS to form condensates under conductance between Col-0 and hln1 mutant became
osmotic stress. Through computational prediction and evident four days after water withholding (Fig. 1B).
experimental analyses, we identified GAD2 as an HLN1 After 10 days of drought, stomatal conductance (mmol
target. GAD2 encodes the primary enzyme for the bio- ­m−2 ­s−1) decreased to 85.17 (Col-0), 119.60 (hln1),
synthesis of γ-aminobutyric acid (GABA), a signaling 85.23(pHLN1-HLN1/hln1 #1), and 87.67 (pHLN1-
molecule implicating in stomatal regulation (Ramesh HLN1/hln1 #2), respectively. The elevated stomatal
et al. 2015; Mekonnen et al. 2016; Xu et al. 2021; Li et al. conductance in hln1 indicated impaired stomatal reg-
2024). We found that HLN1 binds to GAD2 mRNA, pro- ulation. Under the 10-day drought treatment, the sto-
moting condensate formation and transcript stabiliza- matal aperture (width-to-length ratio) in Col-0 plants
tion. Consequently, GAD2 transcripts degraded more decreased from 0.83 to 0.64 but it only decreased from
rapidly in hln1 mutant than in the wild type, leading to 0.84 to 0.78 in the hln1 mutant (Fig. 1C-D). Similar to
reduced GABA levels. Conversely, HLN1 overexpression Col-0, the stomatal aperture in pHLN1-HLN1/hln1 #1
enhanced GAD2 mRNA stability. Consistent with this, and #2 plants decreased from 0.84 to 0.66 and from
Liu et al. Stress Biology (2025) 5:39 Page 3 of 17

0.83 to 0.67, respectively (Fig. 1D). These results indi- results demonstrate that overexpressing HLN1 signifi-
cated that the loss of HLN1 function led to incomplete cantly improves drought tolerance of plants.
stomatal closure and excessive transpiration under
drought. Consequently, the hln1 mutant showed signif- Expression pattern and subcellular localization of HLN1
icantly lower relative water content (RWC) than Col-0, To examine the expression pattern of HLN1, we gen-
while complementation lines restored RWC to wild- erated transgenic Arabidopsis plants expressing the
type levels (Fig. 1E). We also measured stomatal den- β-glucuronidase (GUS) reporter gene driven by a 1,677-
sity but found no significant difference between hln1 bp HLN1 promoter fragment and performed histochemi-
and Col-0 (Supplemental Figs. 1A-1B), suggesting that cal staining. In 7-day-old seedlings, GUS signals were
HLN1 does not regulate stomatal development. detected in cotyledons, lateral roots, root tips, guard cells
Drought stress induces the accumulation of reactive and trichomes (Fig. 2A-F). In 3-week-old plants, GUS
oxygen species (ROS) and malondialdehyde (MDA), a signals were observed in inflorescences, rosette leaves,
product of membrane lipid peroxidation. We observed flowers, and siliques (Fig. 2G-I). Quantitative RT-PCR
significantly higher H­ 2O2 content in the hln1 mutant (qRT-PCR) analysis confirmed higher HLN1 expres-
compared to the wild type, while both hln1 complemen- sion in roots and leaves compared to stems and siliques
tation lines showed reduced levels (Fig. 1F). Addition- (Supplemental Fig. 4). Furthermore, HLN1 expression in
ally, MDA levels were significantly elevated in the hln1 3-week-old plants was significantly enhanced following
mutant (Fig. 1G), indicating more severe lipid peroxida- 10 days of water-withholding (Supplemental Fig. 5).
tion. The complementation lines accumulated less MDA To determine the subcellular localization of the HLN1
than the hln1 mutant, demonstrating that HLN1 loss-of- protein, we generated transgenic lines expressing N-ter-
function exacerbates drought-induced cellular damage. minally-tagged EYFP-HLN1 and C-terminally-tagged
To investigate the effects of HLN1 overexpression on HLN1-EYFP under the control of the 35S promoter.
drought response, we generated transgenic lines express- Roots of the transgenic lines were examined using a con-
ing HLN1 under the control of the cauliflower mosaic focal microscope, and HLN1 was found to be localized in
virus 35S promoter in the Col-0 background. Two inde- the cytoplasm in both EYFP-tagged lines (Fig. 2K).
pendent overexpression lines, both showing over 60-fold
higher HLN1 expression than the wild type (Supplemen- HLN1 undergoes liquid–liquid phase separation and forms
tal Fig. 2), displayed normal growth and development condensates under osmotic stress
under well-water conditions (Supplemental Fig. 3 A), While examining the localization of the EYFP-tagged
indicating that HLN1 overexpression did not impair plant HLN1 protein, we occasionally observed cytoplasmic
growth. When subjected to 23 days of drought treatment condensate-like structures. Structural analysis revealed
followed by re-watering, HLN1-overexpressing plants three potential intrinsically disordered regions (IDRs) in
exhibited enhanced drought tolerance (Supplemental HLN1 (Supplementary Fig. 6), which may drive liquid–
Fig. 3 A). Under the drought treatment, HLN1-overex- liquid phase separation (LLPS) under cellular crowding
pressing plants showed smaller stomatal aperture (Sup- conditions, leading to membraneless condensate for-
plemental Figs. 3D-3E), reduced stomatal conductance, mation. To investigate the phase-separating capability
higher RWC, lower ­H2O2 accumulation, and decreased of the HLN1 protein, we treated 7-day-old YFP-HLN1
MDA levels (Supplemental Figs. 3B-C and H-I). These seedlings with 0.5 M mannitol to induce osmotic stress.

(See figure on next page.)

Fig. 1 HLN1 is required for drought stress response and tolerance in Arabidopsis. A Morphology of Col-0, hln1, complementation line #1 and #2
(Comp #1 and #2) plants before (upper panel) and after (middle panel) a 20-day drought stress, and 2 days after rewatering (lower panel). B
Stomatal conductance in rosette leaves of the indicated plants following drought treatment. Measurements were taken every two days, and data
are means ± standard deviations (SD) from three biological replicates. C Stomatal morphology of the indicated genotypes under drought treatment.
Leaves from 3-week-old plants were excised after 10 days of drought treatment and imaged by light microscope. Experiments were repeated
three times with similar results and representative images are shown. Scale bar = 10 μm. D Stomatal aperture (expressed as width/length ratio)
in leaves of the indicated plants under drought stress. Experiments were repeated three times and stomatal apertures from more than 100 stomata
were calculated and shown as dots. Double asterisks (**) represent a p-value < 0.01 by Student’s t-test. E Relative water content (RWC) in rosette
leaves of the indicated plants after 10-day drought treatment. Data are means ± SD from three biological replicates. Asterisks (*) indicate a p-value
< 0.05 by Student’s t-test. F Hydrogen peroxide ­(H2O2) content in drought-treated plants compared to well-watered controls. Data are means ± SD
from three biological replicates. Asterisks (*) indicate a p-value < 0.05 by Student’s t-test. G Malondialdehyde (MDA) content in drought-treated
plants compared with well-watered plants. Data are means ± SD from three biological replicates. Double asterisks (**) represent a p-value < 0.01
by Student’s t-test. Plant genotypes: Col-0 (wild type); hln1 (hln1 mutant); Comp #1 and #2, two independent complementation lines (i.e.,
pHLN1-HLN1/hln1 #1 and #2)
Liu et al. Stress Biology (2025) 5:39 Page 4 of 17

Fig. 1 (See legend on previous page.)

Liu et al. Stress Biology (2025) 5:39 Page 5 of 17

Fig. 2 Expression pattern and subcellular localization of HLN1. A-F pHLN1:GUS expression in 7-day-old seedling (A), cotyledons (B), lateral roots
(C), root tip (D), guard cells (E), and trichomes (F). G-J pHLN1:GUS expression in the inflorescence (G), rosette leaf (H), flower (I), and silique (J)
of 3-week-old plants. Scale bars: 1000 μm (A-D, G-J); 100 μm (E–F). K Subcellular localization of HLN1 in the root tip cells of pGWB 541-HLN1
and pGWB 542-HLN1 transgenic plants. Scale bar in the lower panel (K) represents 20 μm
Liu et al. Stress Biology (2025) 5:39 Page 6 of 17

Fig. 3 HLN1 forms cytoplasmic condensates via phase-separation both in vivo and in vitro. A Fluorescence Recovery After Photobleaching (FRAP)
analysis of a HLN1 condensate (arrowhead) in the root elongation zone of an EYFP-HLN1 transgenic seedling. Scale bar = 20 μm. B FRAP recovery
kinetics of EYFP-HLN1 condensates in transgenic seedlings. The half-time recovery (­ t1/2) was calculated from averaged fluorescence intensity.
Error bars represent standard deviations (SD) from 9 biological replicates. C in vitro phase separation of HIS-EYFP and HIS-EYFP-HLN1 proteins
following the addition of PEG 8000. Scale bar = 20 μm. D Concentration-dependent formation of HLN1 condensates with PEG 8000. Scale bar = 20
μm. E FRAP analysis of a representative HIS-EYFP-HLN1 droplet. scale bar = 20 μm. F FRAP recovery plot of HIS-EYFP-HLN1 droplets. Half-time
recovery (t₁/₂) was calculated from averaged intensities. Error bars represent SD from 9 biological replicates

(See figure on next page.)

Fig. 4 HLN1 interacts with GAD2 mRNA to regulate its stability in Arabidopsis. A Subcellular localization of HLN1 in root tip cells of transgenic
lines under normal conditions, mannitol treatment, and mannitol treatment combined with cycloheximide (CHX) treatment. Scale bar = 10 μm. B
RIP-qPCR analysis of EYFP-HLN1 binding to GAD2 mRNA with or without dehydration treatment. EYFP-EYFP transgenic line was used as a negative
control. Data are means ± SD (n = 3 biological replicates). **, p < 0.01 by Student’s t-test. C EMSA of HLN1 binding to GAD2 3’UTR analyzed by native
PAGE. MBP-GST was negative control. D GAD2 mRNA decay kinetics in hln1 mutants. Data points show means ± SD (n = 3). E GAD2 mRNA decay
kinetics in HLN1 overexpression (OE) lines. Two independent lines showed similar results. Data from one line shown (means ± SD, n = 3 per time
point). F Relative expression of GAD2 mRNA in Col-0, hln1, and two complementation lines (Comp #1 and #2) under drought stress (means ± SD, n =
3). **, p < 0.01 by Student’s t-test. G GABA content in Col-0, hln1 and two complementation lines (Comp #1 and #2) under control or drought stress
conditions (mean ± SD, n = 3). **, p < 0.01 by Student’s t-test
Liu et al. Stress Biology (2025) 5:39 Page 7 of 17

Fig. 4 (See legend on previous page.)

Liu et al. Stress Biology (2025) 5:39 Page 8 of 17

Fig. 5 Exogenous GABA treatment reduces stomatal conductance and drought-induced damage in the hln1 mutant. A Stomatal morphology
of the indicated genotypes under drought or GABA treatment. Representative results of three independent experiments are shown. Scale bar = 10
μm. B Stomatal apertures (expressed as width/length ratio) of the indicated genotypes with or without drought or GABA treatment. Experiments
were repeated three times and stomatal apertures from more than 100 stomata were calculated. C Stomatal conductance in 3-week-old plants
of the indicated genotypes under drought or 4 mM GABA treatment (means ± SD, n = 3). D Malondialdehyde (MDA) content in 3-week-old plants
of the indicated genotypes under drought or 4 mM GABA treatment (means ± SD, n = 3). E Hydrogen peroxide ­(H2O2) content in 3-week-old plants
of the indicated genotypes under drought or 4 mM GABA treatment (means ± SD, n = 3). Genotypes include Col-0 (wild type); hln1(hln1 mutant);
Comp #1 and #2 (two independent hln1 complementation lines, i.e., pHLN1-HLN1/hln1 #1 and #2). Error bars represent standard deviation (SD). The
double asterisks (**) indicate a p-value < 0.01 and the single asterisk (*) indicates a p-value < 0.05 by Student’s t-test

While EYFP-HLN1 showed even cytosolic distribution mRNA-protein (mRNP) condensates similar to stress
under control conditions, mannitol treatment induced granules, whose formation is CHX-sensitive (Bounedjah
punctuate EYFP-HLN1 aggregates (see below, Fig. 4A). et al. 2014).
To further determine the nature and properties of these We assessed condensate dynamics in the root elon-
HLN1-containing granules, we incubated 7-day-old gation zone using fluorescence recovery after pho-
EYFP-HLN1 seedlings with cycloheximide (CHX) before tobleaching (FRAP). After photobleaching, EYFP-HLN1
the mannitol treatment. It was found that the formation condensates rapidly recovered (Fig. 3A). Quantifica-
of the HLN1 granules was inhibited by CHX treatment tion of the fluorescence intensity showed that the half
(Fig. 4A), suggesting that these cytoplasmic granules are recovery time ­(T1/2) was 43.33 s (Fig. 3B). These data
Liu et al. Stress Biology (2025) 5:39 Page 9 of 17

Fig. 6 Overexpression of GAD2 enhances drought tolerance of the hln1 mutant. A Morphology of 3-week-old plants of the indicated genotypes
before (upper panel) and after (middle panel) 20-day drought stress treatment, and 2 days after rewatering (lower panel). B Stomatal conductance
in rosette leaves after drought treatment. C Relative water content (RWC) in rosette leaves after 10-day drought treatment. D Stomatal morphology
in rosette leaves under drought treatment. Leaves were excised and imaged immediately with a light microscope after 10-day drought treatment.
More than 100 stomata of each genotype were measured, and the experiments were repeated three times with similar results. Scale bar = 10 μm.
E Stomatal apertures (expressed as width/length ratio) under control or drought treatment. Experiments were repeated three times and stomatal
apertures from more than 100 stomata were calculated. F Hydrogen peroxide ­(H2O2) content in rosette leaves under control or drought treatment.
G Malondialdehyde (MDA) content in rosette leaves under control or drought treatment. H GABA content in rosette leaves under control or drought
treatment. Plant genotypes include Col-0 (wild type); hln1 (hln1 mutant); GAD2/hln1 OE #1 and #2 (two independent hln1 complementation lines).
Data in (B-C and F-G) are means and standard deviation (SD) from three biological replicates. The double asterisks (**) indicate a p-value < 0.01,
and the single asterisk (*) indicates a p-value < 0.05 by Student’s t-test

demonstrate that the EYFP-HLN1 protein could undergo HIS-EYFP-HLN1 protein, but not HIS-EYFP, formed
LLPS under osmotic stress to form dynamic condensates spherical droplets (Fig. 3C). The size of HIS-EYFP-HLN1
in vivo. droplets increased in a HLN1 concentration-depend-
To investigate whether HLN1 also undergoes phase ent manner (Fig. 3D). The FRAP assay further revealed
separation in vitro, we purified HIS-EYFP-HLN1 and that the EYFP signals from the HIS-EYFP-HLN1 drop-
HIS-EYFP proteins. Under osmotic stress induced let rapidly recovered after photobleaching (Fig. 3E)
by polyethylene glycol 8000 (PEG8000), only the with a T
­ 1/2 of 23.33 s (Fig. 3F). These data indicate that
Liu et al. Stress Biology (2025) 5:39 Page 10 of 17

HIS-EYFP-HLN1 molecules can diffuse freely within the include ‘mRNA surveillance pathway’, ‘DNA replica-
droplets. Taken together, our in vivo and in vitro assays tion’ and ‘Aminoacyl-tRNA biosynthesis’ (Supplemental
indicate that the HLN1 protein undergoes LLPS to form Fig. 7B).
cytoplasmic granules under osmotic stress conditions. While the biological relevance of these pathways in
hln1 remains unclear, we focused on high-probability
Reduced GAD2 levels and GABA content in hln1 individual transcripts that may interact with HLN1. By
under drought stress integrating prediction scores with single cell expres-
The severe drought-sensitive phenotype of the hln1 sion data from Arabidopsis Leaf Time-Dependent Atlas
mutant suggests that HLN1 regulates key processes in (AraLeTA) and leaf transcriptomics (Vong et al. 2024;
plant drought responses. As an mRNA-binding pro- Tenorio et al. 2025), we identified GAD2 as a top candi-
tein, HLN1 likely influences the metabolism of specific date (Supplemental Figs. 7 C and 7D). GAD2 encodes one
mRNAs critical for stomatal closure during drought of the glutamate decarboxylases that catalyze the decar-
stress. To identify the RNA targets, we conducted RIP- boxylation of glutamate to generate γ-aminobutyric acid
seq (RNA-binding protein immunoprecipitation-RNA (GABA), a non-proteinogenic amino acid known to regu-
sequencing) assays, yet the library construction was late stomatal movement (Mekonnen et al. 2016; Xu et al.
unsuccessful for unclear reasons. We thus resorted 2021).
to alternative approaches to identify potential HLN1 In Arabidopsis, the glutamate decarboxylase (GAD)
targets. family comprises five members. The affinity score of
We performed an in-silico analysis of potential RNA GAD2 to HLN1 was the highest among members (Sup-
targets of HLN1 using RNA–Protein Interaction Pre- plemental Fig. 7 F). In leaf tissues, GAD2 was the most
diction (RPISeq) (Muppirala et al. 2011). This analysis abundant, being at least 40-fold higher than other GAD
predicted hundreds of HLN1-binding mRNAs. Gene members (Supplemental Fig. 7G). This is consistent with
Ontology (GO) enrichment analysis showed that these a previous report identifying GAD2 as the predominant
predicted targets were primarily associated with ‘mem- GAD transcript in Arabidopsis leaves (Miyashita and
brane-bounded organelle (GO:0043227)’, ‘intracellular Good 2008). Under drought treatment, we found that
membrane-bounded organelle (GO:0043231)’ and ‘intra- transcript levels of GAD2 in hln1 mutants were only 63%
cellular organelle (GO:0043229)’ (Supplemental Fig. 7 A). of wild-type levels (see below, Fig. 4F), suggesting that
Kyoto Encyclopedia of Genes and Genomes (KEGG) HLN1 maintains GAD2 expression during drought stress.
enrichment analysis showed that the top pathways Consistent with this finding, hln1 mutants accumulated

Fig. 7 HLN1 condensates stabilize GAD2 mRNA during drought stress. Glutamate decarboxylase 2 (GAD2) catalyzes the conversion of glutamate
to γ-aminobutyric acid (GABA) in leaves. Under well-watered conditions, GAD2 transcript levels remain low, supporting basal GABA synthesis. During
drought stress, GAD2 expression increases. HLN1 binds with GAD2 mRNA and other transcripts and undergoes liquid–liquid phase separation (LLPS)
to form the mRNA-protein (mRNP) condensates that protect GAD2 mRNA from rapid degradation. This stabilization enables sustained and elevated
GAD2 protein production, enhancing GABA synthesis under drought conditions. In the absence of HLN1, GAD2 mRNA becomes less stable,
reducing GABA levels and impairing stomatal closure during drought stress
Liu et al. Stress Biology (2025) 5:39 Page 11 of 17

significantly less GABA than the wild type under drought expression level in hln1 was significantly lower than in
stress (Supplemental Figs. 7E). the wild type (Fig. 4F). Meanwhile, the transcript levels of
Since abscisic acid (ABA) synthesized under abiotic GAD2 in the two complementation lines were restored to
stress triggers stomatal closure, we also examined the levels comparable to the wild type. Furthermore, HLN1-
response of hln1 towards ABA. While ABA induced the overexpression lines showed elevated GAD2 transcript
closure of stomata in both the wild type and the hln1 levels under drought conditions (Supplemental Figs. 3 F
mutant, there were no significant differences in stomatal and 3G).
apertures between them (Supplemental Figs. 8 A and 8B). GAD2 is a key enzyme for GABA synthesis in leaves
Furthermore, drought-induced ABA levels were similar (Miyashita and Good 2008) and altered GAD2 mRNA
between the hln1 mutant and the wild type (Supplemen- levels would affect plant GABA content (Mekonnen et al.
tal Fig. 8 C). Thus, the impaired stomatal closure of hln1 2016). We quantified GABA levels across genotypes.
under drought stress may not be caused by ABA biosyn- While all lines showed similar baseline GABA content
thesis or signaling defects. under control conditions, significant differences were
found under drought stress. hln1 mutants accumulated
HLN1 binds and stabilizes GAD2 transcripts
less GABA than wild-type plants, while both comple-
Based on the above findings, we hypothesized that HLN1 mentation lines restored GABA production to wild-type
regulates drought responses through post-transcriptional levels (Fig. 4G). HLN1 overexpression plants exhibited
modulation of GAD2 mRNA levels. To test this hypoth- even high GABA accumulation (Supplemental Fig. 3G).
esis, we first investigated HLN1–GAD2 interactions Thus, there is a clear correlation among HLN1 dosage,
using both in vivo and in vitro approaches. RIP-qPCR GAD2 transcript levels, and GABA accumulation under
assays performed with EYFP-HLN1 transgenic seedlings drought stress. These results establish that HLN1 stabi-
demonstrated that HLN1 preferentially bound to GAD2 lizes GAD2 transcripts through direct binding and leads
transcripts over EYFP transcripts under normal condi- to enhanced GABA production during drought stress.
tions, with significantly enhanced enrichment of GAD2 The formation of HLN1-GAD2 mRNA condensates likely
transcripts following dehydration treatment (40% fresh represents a key mechanistic step in GABA-mediated
weight loss) (Fig. 4B). Electrophoretic mobility shift stomatal responses to drought stress.
assay (EMSA) confirmed direct interaction, showing spe-
cific binding of MBP-HLN1 to the GAD2 3’UTR​, while
the control protein MBP-GST showed no binding activ- HLN1 regulates stomatal movement through GABA
ity (Fig. 4C). Interestingly, in vitro assays revealed that signaling
GAD2 mRNA promoted HLN1 condensate formation Recent studies have established GABA as a signaling mol-
but failed to form condensates with the control YFP pro- ecule that triggers stomatal closure during drought stress
tein (Supplemental Fig. 9). (Mekonnen et al. 2016; Xu et al. 2021, 2024). To test
To evaluate the functional consequences of the inter- whether HLN1 enhances drought tolerance by regulating
action between HLN1 and GAD2 transcripts, we exam- GAD2 mRNA stability and consequently modulating sto-
ined the GAD2 mRNA stability in the hln1 mutant matal movement through GABA signaling, we examined
background. Seven-day-old seedlings were treated with the effect of exogenous GABA on stomatal responses in
cordycepin to inhibit transcription, and the dynamics the hln1 mutant. Three-week-old plants of Col-0, hln1,
of GAD2 mRNA was monitored over time. It was found and two HLN1 complementation lines were sprayed with
that GAD2 mRNA levels decreased much faster in the 4 mM GABA and subjected to 10 days of water withhold-
hln1 mutant than in the wild type (Fig. 4D). Conversely, ing before stomatal aperture measurement.
GAD2 mRNA in the two independent HLN1 overex- Under well-watered conditions, GABA treatment
pression lines were more stable than in the wild type showed no significant effect on stomatal movement.
(Fig. 4E). These data indicate that HLN1 could stabilize However, during drought stress, we observed two key
GAD2 mRNA, protecting the transcripts from rapid findings: 1) wild type plants showed reduced stomatal
degradation. apertures that were further enhanced by GABA appli-
We measured the expression levels of GAD2 in two cation (Fig. 5A-B), and 2) exogenous GABA completely
independent complementation lines generated by trans- rescued the impaired stomatal closure phenotype of hln1
forming the hln1 mutant with the wild type HLN1 mutants (Fig. 5A-B). These observations are consistent
genomic DNA under the control of its native promoter with previous reports that GABA promotes stomatal clo-
(Sajeev et al. 2022). Under control conditions, the expres- sure and enhances drought tolerance (Xu et al. 2021) and
sion level of GAD2 was similarly low among all geno- indicate that HLN1 regulates stomatal movement at least
types. Drought stress increased GAD2 expression, yet the partially through GABA-mediated pathways.
Liu et al. Stress Biology (2025) 5:39 Page 12 of 17

We further characterized the physiological effects HLN1/AtRGGC/AT5G47210) (Ambrosone et al. 2015;

of GABA treatment by measuring stomatal conduct- Sajeev et al. 2022; Bleckmann et al. 2023). Additionally,
ance, reactive oxygen species (ROS) accumulation, and HLN1 possesses multiple intrinsically disordered regions
lipid peroxidation levels. GABA application significantly (IDRs) that drive LLPS (Hentze et al. 2018). Our assays
reduced all three parameters in hln1 mutants compared confirmed HLN1’s ability to form dynamic, liquid-like
to untreated controls (Fig. 5C-E). These data demonstrate membraneless condensates under conditions such as
that these drought-stress related phenotypic defects in higher osmolarity and cellular crowding (Fig. 3A and C;
the hln1 mutant can be rescued by exogenous GABA Supplemental Fig. 9). This behavior is consistent with
treatment. other RNA-binding proteins (Alshareedah et al. 2020;
Zhu et al. 2022; Fan et al. 2024; Wadsworth et al. 2024;
Overexpression of GAD2 rescues drought sensitivity Wang et al. 2024) and suggests that HLN1’s functions are
in hln1mutants likely mediated by these dynamic RNP condensates.
To determine whether HLN1-sustained GAD2 transcript Previous reports have indicated that AtRGGA
stability contributes to drought tolerance, we introduce a enhances salt and drought tolerance (Ambrosone et al.
35S:Flag-GAD2 construct into hln1 mutants and assessed 2015), though the mechanisms were unknown. Our
their drought responses. As shown in Fig. 6A, overex- analyses identified GAD2 as a candidate mRNA target
pressing GAD2 effectively rescued the drought-induced of HLN1, confirmed by EMSA and RIP-qPCR assays.
seedling lethality observed in hln1 mutants. Notably, Notably, GAD2 is the most abundant GAD family mem-
both GAD2/hln1 overexpressing (OE) lines accumulated ber in Arabidopsis leaves that controls GABA production
endogenous GABA at levels more than double those of (Miyashita and Good 2008). HLN1-GAD2 condensates
the wild type plants (Fig. 6). likely stabilize GAD2 transcripts, leading to higher GABA
Compared with the hln1 mutant, the stomatal conduct- levels under drought conditions. This functionality of
ance and stomatal aperture in both GAD2/hln1 OE lines HLN1 is similar to that of rice DRG9, which stabilizes
decreased under drought stress (Fig. 6B and D-E), leading OsNCED4 mRNA to enhance ABA biosynthesis and
to the restoration of RWC to wild-type levels (Fig. 6C). drought tolerance (Wang et al. 2024). Our pharmacologi-
Consistently, relative to the hln1 mutant, the content cal inhibition experiments demonstrated faster decay of
of ­H2O2 and MDA was reduced in the GAD2/hln1 OE GAD2 mRNA in hln1 mutants and increased stability
lines (Fig. 6F-G), suggesting less oxidative stress in these with HLN1 overexpression. This stabilization correlates
plants under drought conditions. These findings demon- with enhanced drought tolerance, as evidenced by the
strate that GAD2 overexpression compensates for HLN1 recovery of drought tolerance in hln1 mutants through
loss-of-function and support our model in which HLN1 GAD2 overexpression or GABA application (Fig. 6A).
enhances drought tolerance through stabilization of In recent years, GABA has been found to regulate sto-
GAD2 transcripts thereby maintaining adequate GABA matal movement under drought stress (Mekonnen et al.
levels for proper drought responses and tolerance. 2016; Xu et al. 2021), in addition to its diverse functions
in biotic and abiotic stress responses (Islam et al. 2024).
Disscussion GABA interacts with aluminum-activated malate trans-
Recent studies have identified the Arabidopsis HLN1 porters (ALMTs) and modulates the malate uptake into
protein as part of an RNA-binding proteome that inhib- the vacuoles of guard cells, which blocks the opening of
its seed germination through mechanisms that remain stomata and enhances drought tolerance (Ramesh et al.
unclear (Sajeev et al. 2022). Our current research demon- 2015; Gilliham and Tyerman 2016; Xu et al. 2021). Upon
strates HLN1’s critical role in plant drought stress toler- GABA deficiency or depletion under drought stress,
ance. Under osmotic stress conditions, HLN1 undergoes the stomatal closure would be impaired. This is consist-
liquid–liquid phase separation (LLPS) to form cytoplas- ent with our data that loss of function in HLN1 leads to
mic condensates, interacting with stress-related RNAs reduced GABA contents and greater stomatal aperture
such as GAD2. These ribonucleoprotein (RNP) conden- under drought stress (Fig. 1D-E). As the primary enzyme
sates may stabilize GAD2 transcripts, thereby increasing for GABA synthesis in leaves, both the transcriptional
GABA levels and enhancing drought tolerance. and posttranscriptional regulation of GAD2 would be
As an RNA-binding protein (RBP), HLN1 contains important for GABA production.
a conserved HABP4 domain and an RGG repeat motif, We propose a model where HLN1 enhances drought
both are recognized RNA-binding regions (Thanda- tolerance by stabilizing GAD2 mRNA, thereby promoting
pani et al. 2013; Hentze et al. 2018) and are conserved GABA synthesis and facilitating stomatal closure through
across the three-membered Arabidopsis HLN1 fam- interacting with ALMTs (Fig. 7). This mechanism appears
ily (AtRGGA/At4 g16830, AtRGGB/AT4G17520, and to be crucial for maintaining GABA levels and stomatal
Liu et al. Stress Biology (2025) 5:39 Page 13 of 17

regulation during drought stress. While transcriptional plant adaptation to environmental stresses and identify
regulation of stress genes is well-documented, the role new targets for improving stress tolerance.
of posttranscriptional regulation including mRNA sta-
bilization via RBPs is less explored. Previous transgenic Materials and methods
approaches to regulate plant stress resistance are also Plant materials and growth conditions
more focused on regulation of transcription through The Arabidopsis thaliana ecotype Col-0, referred to as
transcription factor manipulations, yet these approaches the wild type, was used in this study. The hln1 knockout
often result in reduced or stunt growth of the transgenic mutant (SALK_055953) and its complementation lines
plants (Gong et al. 2020). In our current study, we did were generously provided by Professor Leonie Bent-
not notice any obvious phenotypic changes of transgenic sink (Sajeev et al. 2022). For growth assays, seeds were
plants overexpressing HLN1 under normal conditions surface-sterilized in 75% ethanol for 10 min and then
(Supplementary Fig. 3). Thus, targeting mRNA stability rinsed five times with sterile Milli-Q water. The sterilized
through RBPs like HLN1 presents a promising strategy seeds were sown on ½-strength Murashige and Skoog
for engineering drought-resistant crops without yield (MS) medium plates containing 1% agar and 1% sucrose.
penalties. The plates were incubated at 4 ℃ for 3 days before being
transferred to a plant growth room maintained at 22 ℃
Limitations of the study with a 16-h light/8-h dark cycle.
While our study provides insights into HLN1’s role in
drought tolerance, the current work has several limita- Drought treatment and relative water content (RWC)
tions. First, the identification of HLN1’s full mRNA tar- measurement
get repertoire remains incomplete. While GAD2 was Seven-day-old seedlings were transferred to pots con-
confirmed as a key target, HLN1 likely binds other tran- taining a uniform amount of soil (250 g) with 9 seedlings
scripts that contribute to drought tolerance or other per pot. For drought treatment, 3-week-old plants were
physiological processes. Second, the precise roles of subjected to a 20-day period without watering, followed
HLN1’s IDRs in condensate formation and RNA binding by rewatering. The morphology of the plants was docu-
are unclear. Our initial mutational analyses of individual mented using a Canon EOS R7 camera at the beginning
IDR functions in RNA-binding and condensate forma- and at the end of the drought treatment and at 2 days
tion were challenging and more sophisticated approaches after rewatering.
to dissect IDR functionality are needed. Third, this study Relative water content (RWC) was calculated using the
focused primarily on GABA-mediated stomatal regula- following equation: RWC = (FW-DW)/(SW-DW), where
tion, but HLN1’s role in other drought-related processes, FW represents fresh weight, DW represents the dry
such as reactive oxygen species (ROS) suppression weight and SW represents saturated weight.
(Li et al. 2021; Xu et al. 2024), remains unexplored.
Finally, while our proposed model emphasizes transcript Stomatal aperture, conductance, and density
stability, HLN1 may also regulate transcript processing For stomatal aperture analysis, the fifth or sixth pair of
and translation, as suggested by its homologs in other rosette leaves from 3-week-old plants was excised, and
species (Ma et al. 2020). These processes could poten- epidermal peels were immediately prepared for imaging
tially impact GAD2 protein levels or enzymatic activities under a light microscope equipped with a digital camera.
and warrant further investigation in order to fully under- Stomatal apertures were measured using ImageJ software
stand HLN1’s role in stress adaptation. (NIH, USA).
Stomatal conductance was measured using a METER
Conclusions SC-1 Leaf Porometer (Team Medical & Scientific Sdn.
Our study establishes HLN1 as a critical drought toler- Bhd., Malaysia). The instrument was first calibrated
ance regulator that stabilizes GAD2 mRNA, enhancing according to the manufacturer’s instructions prior to
GABA production, and regulating stomatal closure dur- measurements. The sensor was clipped onto the leaves,
ing drought stress. This research provides insights into and readings were recorded after stabilization.
posttranscriptional regulation of stress genes and under- Stomatal density was determined by photographing
scores the significance of RBPs in plant stress responses. leaf epidermis peels. Specifically, the fifth or sixth pair of
Targeting mRNA stability through RBPs like HLN1 offers rosette leaves from 3-week-old plants was excised, and
promising strategies for engineering drought-resistant epidermal peels were immediately mounted on micro-
crops without compromising yield. Further studies on scope slide for observation under a light microscope
HLN1 and other RBPs may reveal novel mechanisms of equipped with a digital camera. Stomata were counted
Liu et al. Stress Biology (2025) 5:39 Page 14 of 17

under each documented view to determine density destination vectors, including pEarleyGate 101-HLN1
(number/mm2). or pGWB541-HLN1 for C-terminal tagging and pEarl-

tagging, were generated via Gateway™ LR Clonase™ II

eyGate 104-HLN1 or pGWB542-HLN1 for N-terminal
Hydrogen peroxide, MDA and GABA quantification
Leaf hydrogen peroxide (­ H2O2) of 3-week-old plants was Enzyme mix (Cat. # 11,791,020, Thermo). For GAD2

the pDONR™/Zeo vector. The destination vector pEar-

assayed using a Hydrogen Peroxide Assay Kit (BC3595, overexpression in hln1, GAD2 cDNA was cloned into
Solarbio Co., Beijing, China) following the manufactur-
er’s instruction. leyGate 202-GAD2 was obtained through LR recom-
Malondialdehyde (MDA) content was measured using bination. Transgenic plants were generated using
the method described by Hodges et al. (1999). Briefly, Agrobacterium tumefaciens (strain GV3101) -mediated
0.5 g of fresh plant samples was homogenized in 2 mL floral dip transformation. All primers are listed in Sup-
of chilled 10% TCA (trichloroacetic acid) buffer. After plemental Table 2.
centrifugation at 10,000 g for 15 min, the supernatant
was mixed with an equal volume of 0.6% thiobarbituric GUS staining
acid (TBA). The mixture was then heated at 100 ℃ for 20 The 1677 bp promoter of HLN1 was amplified from
min, cooled on ice, and centrifuged again at 10,000 g for Arabidopsis genomic DNA and cloned into the pDONR-
10 min. The supernatant was collected and the absorb- HLN1 vector. The pGWB433-pHLN1 vector was
ance measurement at 532 nm, 600 nm and 450 nm were obtained via LR reaction. Transgenic plants were gener-
obtained using a Shimadzu UV-1800 spectrophotometer. ated using the floral dip transformation method. GUS
The MDA content was calculated according to the for- staining was performed with ­T3 generation seedlings or
mula: 6.45 × ­(A532 – ­A600) – 0.56 × ­A450.
GABA content was determined using the Elabscience®
adult plants using GUS Stain solution (BN20175, Biori-
gin) according to the manufacturer’s instructions. Tissues
GABA Colorimetric Assay Kit (E-BC-K852-M, Elabsci- were incubated in the X-gluc solution at room tem-
ence) according to the manufacturer’s instructions. In perature overnight, then destained in ethanol prior to
brief, leaf samples from 3-week-old plants, with or with- observation.
out 10-day water-withheld treatment, were ground with
the extraction solution. The supernatant was reacted with Subcellular localization
phenol and sodium hypochlorite to produce a blue-green Confocal images were acquired using a Leica confo-
product that has a maximum absorbance at 640 nm. The cal microscope. The subcellular localization of HLN1
GABA content was calculated based on absorbance of was determined in root tips from 7-day-old transgenic
samples in comparison with the standard curves. seedlings. Transgenic plants expressing either N-ter-
minal EYFP tagged and C-terminal EYFP tagged HLN1
ABA content and ABA‑induced stomatal closure fusion proteins were examined, and similar results were
ABA content was measured using an Abscisic acid obtained. For mannitol treatment, 7-day-old seedlings
(ABA) ELISA Kit (Fankew Industrial Co., Ltd., Shanghai). were incubated in 0.4 M mannitol solution for 30 min.
Briefly, leaves from 3-week-old Arabidopsis plants, with For CHX treatment, 7-day-old seedlings were first incu-
or without a 10-day drought treatment, were harvested bated with 50 μM CHX for 30 min at room temperature
for ABA quantification following the manufacturer’s pro- and then transferred to 0.4 M mannitol solution for an
tocol. Absorbance was measured using a MULTISKAN additional 30 min before observation.
GO Microplate Reader (Thermo).
For stomatal aperture assays, leaves from 3-week-old Fluorescence recovery after photobleaching (FRAP)
Col-0 and hln1 plants were excised and immersed in a For in vivo FRAP experiments on the EYFP-HLN1 con-
Leaf Open Buffer (10 mM MES-Tris (pH 5.6), 5 mM KCl, densates in root elongation zone cells, we used a Leica
50 μM CaCl2) under light for 2 h, followed by transfer Stellaris 5 confocal microscope with a × 63 objective
to the same containing 25 μM ABA. After 2 h incuba- lens. EYFP-HLN1 condensates were bleached using a
tion, epidermal peels were prepared and imaged using an 488-nm pulse at 50% intensity. Fluorescence recovery
Olympus CKX41 microscope. was recorded every 10 s for up to 160 s post-bleaching.
Images were acquired and quantified using the LAS X
Plant transformation software.

amplified and cloned into the pDONR™/Zeo vector

For HLN1 subcellular localization, HLN1 cDNA was For in vitro FRAP analysis of HIS-EYFP-HLN1 drop-

(Invitrogen) using Gateway™ BP Clonase™ II Enzyme

lets, the central area of selected droplets was bleached
using a 488-nm pulse at 50% intensity. Fluorescence
mix (Cat. # 11,789,020, Thermo). The EYFP-tagged recovery was similarly monitored every 10 s for up to 160
Liu et al. Stress Biology (2025) 5:39 Page 15 of 17

s post-bleaching. To calculate the recovery rate, the fluo- performed, each using approximately 0.5 g of sample.
rescence values were normalized to the first time point Total RNA was isolated for subsequent qRT-PCR analy-
and expressed as a percentage of the pre-bleach intensity. sis, using UBQ5 as the internal control. Primers used for
qRT-PCR are listed in Supplemental Table 1.
Protein expression and purification
To construct the pET28a-EYFP-HLN1 and pET28a-EYFP RIP‑qPCR analysis
plasmids, the EYFP-HLN1 and EYFP fragments were The RIP assay was conducted similarly as described (Song
cloned from the pGWB542-HLN1 plasmid and inserted et al. 2023). Briefly, 14-day-old seedlings were harvested,
into the pET28a plasmid using the ClonExpress II One and approximately 3-g of sample were treated with 30 mL
Step Cloning Kit (C112, Vazyme) at the EcoRI and Hin- of 1% formaldehyde solution under vacuum for 15 min.
dIII sites. For the MBP-HLN1 plasmid, the HLN1 insert The fixation reaction was quenched using 2 M glycine
was cloned into the pMAL-c5x plasmid at the SalI site for 5 min under vacuum. The samples were then washed
via homologous recombination. For prokaryotic expres- three times with water and blotted on paper towels to

(DE3) strain and proteins were purified with Pierce™

sion, the plasmids were expressed in the E. coli BL21 remove excess water. Subsequently, the samples were fro-
zen with liquid nitrogen and ground into powder. Lysis
Glutathione Magnetic Agarose Beads (78,601, Thermo) buffer was added to powdered samples, and the mixture
or Amylose Magnetic Beads (E8035S, NEB) according to was incubated at 4 ℃ for 30 min. After centrifugation,
the manufacturer’s instructions. half of the lysate were saved as input, while the remain-
ing lysate was incubated with anti-GFP magnetic beads
EMSA analysis (P2132, Beyotime Biotech Inc, Shanghai, China) at 4 ℃ for
To generate the 3’UTR and full-length mRNA of GAD2 2 h with rotation. Following beads washing and proteinase
via in vitro transcription, the GAD2 3’UTR and the full- K treatment, RNA was extracted using Trizol reagent. The
length cDNA were synthesized and inserted into the relative enrichment of RNA was determined by qRT-PCR.
pcDNA3.1(+) plasmid driven by a T7 promoter. The For the dehydration treatment, seedlings were exposed to
3’UTR and full-length mRNA transcripts were then syn- air until 40% of their fresh weight was lost. The primers
thesized using the T7 High Yield RNA Transcription Kit used for this assay are listed in Supplemental Table 1.
(TR101-01, Vazyme) and were purified using the Mon-
arch RNA Cleanup Kit (T2040S, NEB). The protein- In silico RIP‑seq analysis
mRNA interaction assays were performed as previously Transcripts from Arabidopsis Col-0 plants were obtained
described (Seo et al. 2019). Briefly, mRNA was incubated from the Ensembl database, and the HLN1 protein
with the MBP-HLN1 protein at room temperature for 30 sequence was retrieved from the NCBI database. These
min and analyzed by native PAGE gel. RNA shift signals sequences were analyzed using the RPIseq web server
were detected using Ultra GelRed Nucleic Acid Stain (GR (http://​pridb.​gdcb.​iasta​te.​edu/​RPISeq/) in batch sub-
501–01, Vazyme). mission mode. A cutoff threshold of > 0.95 for both the
random forest (RF) and support vector machine (SVM)
RNA isolation and real‑time PCR classifiers was applied to identify positive hits. The inter-
Total RNA was extracted from 3-week-old plants using acting candidates were ranked based on gene expression
RNeasy Plant Mini Kit (74,904, Qiagen) along with the data in Arabidopsis leaves (Vong et al. 2024; Tenorio et al.
RNase-Free DNase Set (79,254, Qiagen). cDNA synthe- 2025). Functional annotation of the positive hits was per-
sis was performed using the HiScript II 1 st Strand cDNA formed using Gene Ontology (GO) and KEGG pathway
Synthesis Kit (R211-01, Vazyme). Real-time qPCR reac- analyses via the OmicShare platform (Mu et al. 2024).
tions were performed using the ABI PCR System with
ChamQ Universal SYBR qPCR Master Mix (Q711-02, Functional domain and intrinsically disordered region
Vazyme), using 18S rRNA as the internal control. Three analysis
biological replicates were performed, and the standard The functional domains of HLN1 were analyzed using
deviation (SD) was calculated. Primers used in this study InterPro (Blum et al. 2025). The intrinsically disordered
are listed in Supplemental Table 1. regions (IDRs) of the HLN1 protein were predicted using
PONDR (Peng et al. 2006). Illustrations were generated
mRNA decay analysis using GraphPad.
Seven-day-old seedlings were incubated with 1 mM
Abbreviations
cordycepin (HY-N0262, MCE) and vacuum infiltrated ABA Abscisic acid
for 15 min. After the treatment, samples were harvested ALMT Aluminum-activated malate transporter
at specific time points. Three biological replicates were AraLeTA Arabidopsis Leaf Time-Dependent Atlas
Liu et al. Stress Biology (2025) 5:39 Page 16 of 17

AtRGGA/B/C Arabidopsis RGG (Arg-Gly-Gly) motif protein A/B/C Authors’ contributions

CHX Cycloheximide L.X. and C.L. conceived the project idea. C.L. did the experiments and analyzed
DRG9 Drought resistance gene 9 the data. Y.W., J.P., Z.S., Y.L., Z.Z., X.M., Y.Y. and T.Q. helped experiments and data
DW Dry weight analysis. C.L. wrote the manuscript. L.X., T.Q., and Y.X. revised the manuscript.
EMSA Electrophoretic mobility shift assay
EYFP Enhanced yellow fluorescent protein (YFP) Funding
FRAP Fluorescence recovery after photobleaching This study was supported by Hong Kong University Grants Committee (UGC)
FW Fresh weight General Research Fund (GRF) Grant #12103020 (to L.X) and Collaborative
GABA γ‐Aminobutyric acid Research Fund (CRF) grant #C2003-22 WF (to Y.X and L.X).
GAD1/2 Glutamate decarboxylase 1/2
GFP Green fluorescent protein Data availability
GO Gene Ontology All data supporting the findings of this study are included in the article and its
GUS β-Glucuronidase supplementary materials.
HABP4/SERBP1 Hyaluronan-binding protein 4 / serpin mRNA binding protein 1
HIS Histidine
HLN1 Hyaluronan 1 Declarations
IDR Intrinsically disordered region
KEGG Kyoto Encyclopedia of Genes and Genomes Ethics approval and consent to participate
LLPS Liquid–liquid phase separation Not applicable.
LOT1 Low Temperature 1
MDA Malondialdehyde Consent for publication
mRNP mRNA-ribonucleoprotein All authors have read and approved the manuscript’s content and consent to
MS Murashige and Skoog (salt) its publication in Stress Biology upon acceptance.
OsNCED4 Rice 9-cis-epoxycarotenoid dioxygenase 4
PAGE Polyacrylamide gel electrophoresis Competing interests
P-bodies Processing bodies L.X. is a member of the editorial board but was not involved in the journal’s
PEG Polyethylene glycol review, or any decisions, related to this submission.
qRT-PCR Quantitative RT-PCR
RBP RNA-binding protein
RF Random forest Received: 22 April 2025 Revised: 6 May 2025 Accepted: 13 May 2025
RGG​ Arg-Gly-Gly (motif )
RIP-qPCR RNA-binding protein immunoprecipitation-quantitative RT-PCR
RIP-seq RNA-binding protein immunoprecipitation-RNA sequencing
ROS Reactive oxygen species
RPISeq RNA–Protein Interaction Prediction References
RWC​ Relative water content Alshareedah I, Moosa MM, Raju M, Potoyan DA, Banerjee PR (2020) Phase
SD Standard deviation transition of RNA-protein complexes into ordered hollow condensates.
SVM Support vector machine Proc Natl Acad Sci U S A 117(27):15650–15658. https://​doi.​org/​10.​1073/​
SW (Water) saturated weight pnas.​19223​65117
TBA Thiobarbituric acid Ambrosone A, Batelli G, Nurcato R, Aurilia V, Punzo P, Bangarusamy DK, Ruberti
TCA​ Trichloroacetic acid I, Sassi M, Leone A, Costa A, Grillo S (2015) The Arabidopsis RNA-binding
UBQ5  UBIQUITIN 5 protein AtRGGA regulates tolerance to salt and drought stress. Plant
UTR​ Untranslated region Physiol 168(1):292–306. https://​doi.​org/​10.​1104/​pp.​114.​255802
X-gluc 5-Bromo-4-chloro-3-indolyl-β-D-glucuronide Bleckmann A, Spitzlberger N, Denninger P, Ehrnsberger HF, Wang L, Bruck-
mann A, Reich S, Holzinger P, Medenbach J, Grasser KD, Dresselhaus T
(2023) Cytosolic RGG RNA-binding proteins are temperature sensitive
Supplementary Information flowering time regulators in Arabidopsis. Biol Chem 404(11–12):1069–
The online version contains supplementary material available at https://​doi.​ 1084. https://​doi.​org/​10.​1515/​hsz-​2023-​0171
org/​10.​1007/​s44154-​025-​00239-4. Bounedjah O, Desforges B, Wu TD, Pioche-Durieu C, Marco S, Hamon L, Curmi
PA, Guerquin-Kern JL, Piétrement O, Pastré D (2014) Free mRNA in excess
upon polysome dissociation is a scaffold for protein multimerization to
Supplementary Material 1. Supplemental Figure 1. Stomatal density in
form stress granules. Nucleic Acids Res 42(13):8678–8691. https://​doi.​org/​
rosette leaves of Col-0 and hln1. Supplemental Figure 2. Relative HLN1
10.​1093/​nar/​gku582
expression in HLN1 overexpression lines. Supplemental Figure 3. Over-
Emenecker RJ, Holehouse AS, Strader LC (2020) Emerging Roles for Phase
expression of HLN1 confers drought tolerance. Supplemental Figure 4.
Separation in Plants. Dev Cell 55(1):69–83. https://​doi.​org/​10.​1016/j.​
Relative expression of HLN1 in different tissues. Supplemental Figure 5.
devcel.​2020.​09.​010
Expression of HLN1 was induced by drought stress. Supplemental
Fan S, Zhang Y, Zhu S, Shen L (2024) Plant RNA-binding proteins: Phase separa-
Figure 6. Intrinsically disordered region (IDR) domain analysis of HLN1.
tion dynamics and functional mechanisms underlying plant develop-
Supplemental Figure 7. In-silico analysis of candidate targets of HLN1.
ment and stress responses. Mol Plant 17(4):531–551. https://​doi.​org/​10.​
Supplemental Figure 8. Stomatal response to ABA and ABA content of the
1016/j.​molp.​2024.​02.​016
wild type and hln1 mutant under drought stress. Supplemental Figure 9.
Gilliham M, Tyerman SD (2016) Linking Metabolism to Membrane Signaling:
HLN1 condensates with GAD2 mRNA in vitro. Supplemental Figure 10.
The GABA-Malate Connection. Trends Plant Sci 21(4):295–301. https://​doi.​
Functional domain analysis of HLN1. Supplemental Table 1. Primers used
org/​10.​1016/j.​tplan​ts.​2015.​11.​011
in genotyping and qRT-PCR analysis. Supplemental Table 2. Primers used
Gong Z, Xiong L, Shi H, Yang S, Herrera-Estrella LR, Xu G, Chao DY, Li J, Wang PY,
to construct plasmids.
Qin F, Li J, Ding Y, Shi Y, Wang Y, Yang Y, Guo Y, Zhu JK (2020) Plant abiotic
stress response and nutrient use efficiency. Sci China-Life Sci 63(5):635–
674. https://​doi.​org/​10.​1007/​s11427-​020-​1683-x
Acknowledgements
Hentze MW, Castello A, Schwarzl T, Preiss T (2018) A brave new world of RNA-
We thank Professor Leonie Bentsink for kindly providing the hln1 mutant
binding proteins. Nat Rev Mol Cell Biol 19(5):327–341. https://​doi.​org/​10.​
and its complementation lines. We also thank Lu Sun and Liaoliao Ye for their
1038/​nrm.​2017.​130
advice and experimental assistance.
Liu et al. Stress Biology (2025) 5:39 Page 17 of 17

Hirose T, Ninomiya K, Nakagawa S, Yamazaki T (2023) A guide to membrane- Thandapani P, O’Connor TR, Bailey TL, Richard S (2013) Defining the RGG/RG motif.
less organelles and their various roles in gene regulation. Nat Rev Mol Mol Cell 50(5):613–623. https://​doi.​org/​10.​1016/j.​molcel.​2013.​05.​021
Cell Biol 24(4):288–304. https://​doi.​org/​10.​1038/​s41580-​022-​00558-8 Vong GYW, McCarthy K, Claydon W, Davis SJ, Redmond EJ, Ezer D (2024)
Hodges DM, DeLong JM, Forney CF, Prange RK (1999) Improving the thiobar- AraLeTA: An Arabidopsis leaf expression atlas across diurnal and develop-
bituric acid-reactive-substances assay for estimating lipid peroxidation in mental scales. Plant Physiol 195(3):1941–1953. https://​doi.​org/​10.​1093/​
plant tissues containing anthocyanin and other interfering compounds. plphys/​kiae1​17
Planta 207:604–611. https://​doi.​org/​10.​1007/​s0042​50050​524 Wadsworth GM, Srinivasan S, Lai LB, Datta M, Gopalan V, Banerjee PR (2024)
Islam SNU, Kouser S, Hassan P, Asgher M, Shah AA, Khan NA (2024) Gamma- RNA-driven phase transitions in biomolecular condensates. Mol Cell
aminobutyric acid interactions with phytohormones and its role in 84(19):3692–3705. https://​doi.​org/​10.​1016/j.​molcel.​2024.​09.​005
modulating abiotic and biotic stress in plants. Stress Biol 4(1):36. https://​ Wang H, Ye T, Guo Z, Yao Y, Tu H, Wang P, Zhang Y, Wang Y, Li X, Li B, Xiong H,
doi.​org/​10.​1007/​s44154-​024-​00180-y Lai X, Xiong L (2024) A double-stranded RNA binding protein enhances
Kosmacz M, Gorka M, Schmidt S, Luzarowski M, Moreno JC, Szlachetko J, drought resistance via protein phase separation in rice. Nat Commun
Leniak E, Sokolowska EM, Sofroni K, Schnittger A, Skirycz A (2019) Protein 15(1):2514. https://​doi.​org/​10.​1038/​s41467-​024-​46754-2
and metabolite composition of Arabidopsis stress granules. New Phytol Wiedner HJ, Giudice J (2021) It’s not just a phase: function and characteris-
222(3):1420–1433. https://​doi.​org/​10.​1111/​nph.​15690 tics of RNA-binding proteins in phase separation. Nat Struct Mol Biol
Li L, Dou N, Zhang H, Wu C (2021) The versatile GABA in plants. Plant Signal 28(6):465–473. https://​doi.​org/​10.​1038/​s41594-​021-​00601-w
Behav 16(3):1862565. https://​doi.​org/​10.​1080/​15592​324.​2020.​18625​65 Xu B, Long Y, Feng X, Zhu X, Sai N, Chirkova L, Betts A, Herrmann J, Edwards
Li J, Liu X, Chang S, Chu W, Lin J, Zhou H, Hu Z, Zhang M, Xin M, Yao Y, Guo W, EJ, Okamoto M, Hedrich R, Gilliham M (2021) GABA signalling modu-
Xie X, Peng H, Ni Z, Sun Q, Long Y, Hu Z (2024) The potassium transporter lates stomatal opening to enhance plant water use efficiency and
TaNHX2 interacts with TaGAD1 to promote drought tolerance via modu- drought resilience. Nat Commun 12(1):1952. https://​doi.​org/​10.​1038/​
lating stomatal aperture in wheat. Sci Adv 10(15):eadk4027. https://​doi.​ s41467-​021-​21694-3
org/​10.​1126/​sciadv.​adk40​27 Xu B, Feng X, Piechatzek A, Zhang S, Konrad KR, Kromdijk J, Hedrich R, Gilliham
Ma X, Ibrahim F, Kim EJ, Shaver S, Becker J, Razvi F, Cerny RL, Cerutti H (2020) M (2024) The GABA shunt contributes to ROS homeostasis in guard cells
An ortholog of the Vasa intronic gene is required for small RNA-mediated of Arabidopsis. New Phytol 241(1):73–81. https://​doi.​org/​10.​1111/​nph.​
translation repression in Chlamydomonas reinhardtii. Proc Natl Acad Sci U 19390
S A 117(1):761–770. https://​doi.​org/​10.​1073/​pnas.​19083​56117 Zhu JK (2016) Abiotic Stress Signaling and Responses in Plants. Cell
Marondedze C, Thomas L, Gehring C, Lilley KS (2019) Changes in the Arabi- 167(2):313–324. https://​doi.​org/​10.​1016/j.​cell.​2016.​08.​029
dopsis RNA-binding proteome reveal novel stress response mechanisms. Zhu S, Gu J, Yao J, Li Y, Zhang Z, Xia W, Wang Z, Gui X, Li L, Li D, Zhang H, Liu
BMC Plant Biol 19(1):139. https://​doi.​org/​10.​1186/​s12870-​019-​1750-x C (2022) Liquid-liquid phase separation of RBGD2/4 is required for heat
Mekonnen DW, Flügge UI, Ludewig F (2016) Gamma-aminobutyric acid deple- stress resistance in Arabidopsis. Dev Cell 57(5):583-597.e6. https://​doi.​org/​
tion affects stomata closure and drought tolerance of Arabidopsis thali- 10.​1016/j.​devcel.​2022.​02.​005
ana. Plant Sci 245:25–34. https://​doi.​org/​10.​1016/j.​plant​sci.​2016.​01.​005
Miyashita Y, Good AG (2008) Contribution of the GABA shunt to hypoxia-
induced alanine accumulation in roots of Arabidopsis thaliana. Plant Cell Publisher’s Note
Physiol 49(1):92–102. https://​doi.​org/​10.​1093/​pcp/​pcm171 Springer Nature remains neutral with regard to jurisdictional claims in pub-
Mu H, Chen J, Huang W, Huang G, Deng M, Hong S, Ai P, Gao C, Zhou H (2024) lished maps and institutional affiliations.
OmicShare tools: A zero-code interactive online platform for biological
data analysis and visualization. iMeta 3(5):e228. https://​doi.​org/​10.​1002/​
imt2.​228
Muppirala UK, Honavar VG, Dobbs D (2011) Predicting RNA-protein interac-
tions using only sequence information. BMC Bioinformatics 12:489.
https://​doi.​org/​10.​1186/​1471-​2105-​12-​489
Peng K, Radivojac P, Vucetic S, Dunker AK, Obradovic Z (2006) Length-depend-
ent prediction of protein intrinsic disorder. BMC Bioinformatics 7:208.
https://​doi.​org/​10.​1186/​1471-​2105-7-​208
Qin T, Tian Q, Wang G, Xiong L (2019) LOWER TEMPERATURE 1 Enhances ABA
Responses and Plant Drought Tolerance by Modulating the Stability and
Localization of C2-Domain ABA-Related Proteins in Arabidopsis. Mol Plant
12(9):1243–1258. https://​doi.​org/​10.​1016/j.​molp.​2019.​05.​002
Ramesh SA, Tyerman SD, Xu B, Bose J, Kaur S, Conn V, Domingos P, Ullah S, Wege
S, Shabala S, Feijó JA, Ryan PR, Gilliham M (2015) GABA signalling modu-
lates plant growth by directly regulating the activity of plant-specific anion
transporters. Nat Commun 6:7879. https://​doi.​org/​10.​1038/​ncomm​s8879
Sajeev N, Baral A, America AHP, Willems LAJ, Merret R, Bentsink L (2022) The
mRNA-binding proteome of a critical phase transition during Arabidopsis
seed germination. New Phytol 233(1):251–264. https://​doi.​org/​10.​1111/​
nph.​17800
Sato H, Mizoi J, Shinozaki K, Yamaguchi-Shinozaki K (2024) Complex plant
responses to drought and heat stress under climate change. Plant J
117(6):1873–1892. https://​doi.​org/​10.​1111/​tpj.​16612
Seo M, Lei L, Egli M (2019) Label-Free Electrophoretic Mobility Shift Assay
(EMSA) for Measuring Dissociation Constants of Protein-RNA Complexes.
Curr Protoc Nucleic Acid Chem 76(1):e70. https://​doi.​org/​10.​1002/​cpnc.​70
Song P, Wei L, Chen Z, Cai Z, Lu Q, Wang C, Tian E, Jia G (2023) m6A readers
ECT2/ECT3/ECT4 enhance mRNA stability through direct recruitment
of the poly(A) binding proteins in Arabidopsis. Genome Biol 24(1):103.
https://​doi.​org/​10.​1186/​s13059-​023-​02947-4
Tenorio Berrío R, Verhelst E, Eekhout T, Grones C, De Veylder L, De Rybel B,
Dubois M (2025) Dual and spatially resolved drought responses in the
Arabidopsis leaf mesophyll revealed by single-cell transcriptomics. New
Phytol 246(3):840–858. https://​doi.​org/​10.​1111/​nph.​20446