2n pollen production by extreme low temperature-induced failure of male post-meiotic cytokinesis in Populus Authors: Jiahua Du, Meiling Long, Yunyun Xie, Dongliang Cao, Lexun Ma, Liang Li, Dawei Cheng, Bo Kong, Yifan Zhao, Yaru Sang, Pingdong Zhang Published: 2026 ABSTRACT AbstractPolyploidization by forming diploid gametes is an important force driving diversification and speciation in plant evolution. The prevalence of polyploids in adverse climates led us to underline that abiotic stress conditions can result in unreduced gametes. In the present study, we revealed that short periods of extreme low temperature (ELT) caused defects in male post-meiotic cytokinesis by interfering with the formation of the radial microtubule array, resulting in meiotic restitution and 2n pollen production in Populus. A protocol for inducing 2n pollen by ELT was established. The dominant meiotic stage (F = 19.804, P < 0.001), low temperature (F = 30.656, P < 0.001), and duration (F = 5.554, P = 0.005) significantly affected the frequency of induced 2n pollen. The highest frequency of induced 2n pollen (31.11%) was observed when the dominant meiotic stage of the pollen mother cells was metaphase II and was stressed by -13℃ for 5 h. Twenty-six triploids were produced by crossing induced 2n pollen. Therefore, we developed a reliable and efficient method for triploid breeding of plants. Our findings provide new insight into the mechanisms of ELT-induced 2n pollen formation in a woody plant and verify that sensitivity to environmental stress has evolutionary importance in terms of polyploidization.Keywords: Extreme low temperature, Tubulin immunolocalization, Male meiotic cytokinesis, 2n pollen, Triploid Introduction Polyploidization is one of most important forces driving diversification and speciation in plant evolution [1]. The redundant genomic material created by polyploidization provides genotypic plasticity that facilitates adaptation and improves competitiveness compared with diploid counterparts [1–3]. Molecular evidences indicate that the genomes of most angiosperms (> 90%) retain evidence of one or more ancient genome-wide duplication events [4]. Moreover, Wood et al., [5] established that up to 15% of angiosperm and 31% of gymnosperm speciation events have been accompanied by polyploidization. Although several mechanisms produce polyploids, it is thought that most polyploidy plants are generated by the fusion of 2n gametes [6, 7]. Populus is one of the most widely distributed trees in the world, valued for its fast timbering and short rotation. Triploid breeding of Populus holds considerable theoretical and practical importance. Compared with diploids, triploid Populus combines the advantages of hybrid genomes and polyploidy, exhibiting rapid growth, high basic wood density, long fibers, and high cellulose content [8]. Populus is a model perennial woody plant system. In this system, after two rounds of meiosis, pollen mother cells (PMCs) develop into a tetrad containing four haploid interphase nuclei. Subsequently, male cytokinesis occurs and four unicellular haploid pollen grains are formed. Then, bicellulate pollen with a generative cell and a vegetative nucleus is produced when the unicellular haploid pollen grains undergo mitosis. The PMCs are highly sensitive to changes in the environment during male meiosis. If a physical mutagenic agent is provided, such as high temperature, cold stress, ultraviolet rays, or gamma rays [9–12], meiosis will be inhibited and 2n pollen is generated [13]. Hybridization between unreduced gametes (2n) and normal haploid gametes is a effective method to create triploid plants. However, the natural occurrence of 2n gametes is too low to meet the needs of polyploid breeding. Therefore, artificial induction of 2n gametes by mutagenic agents is required to increase 2n gamete yield and thereby improve the efficiency of triploid production. Low temperature is an important mutagenic agent that induces 2n gametes in higher plants. Some previous studies documented that low temperature has been used to induce 2n pollen during macrosporogenesis in Capsicum annuum, Diospyros kaki, Dioscorea parviflora, Brassica spp., and Arabidopsis thaliana [14–18]. However, owing to the lack of a suitable protocol for inducing 2n pollen by low temperature, a low frequency (< 10%) of induced 2n pollen was obtained with no triploids. Moreover, the mechanism of low temperature-induced 2n pollen formation remains poorly understood. Therefore, this study aimed to investigate the cytological mechanisms by which extreme low temperature (ELT) defects radial microtubule arrays (RMAs) during male meiotic cytokinesis in Populus, resulting in meiotic restitution and 2n pollen production. We established an ELT‑based induction protocol, with the highest 2n pollen frequency (31.11%) achieved at metaphase II under − 13℃ for 5 h, and obtained 26 triploids. Our findings provide new insight into the mechanisms of low temperature-induced 2n pollen formation in a woody plant and verify that sensitivity to environmental stress has evolutionary importance in terms of polyploidization. Materials and methods Plant materials Male floral branches of Populus tomentosa (2n = 2x = 38) clone 145 and female floral branches of P. tomentosa clone 121 (2n = 2x = 38) were collected from a clonal trial in Shuozhou City, Shanxi Province (Forest Tree Seedling Cultivation Center). Male floral branches of P. alba × P. glandulosa ‘84K’ (2n = 2x = 38) and female floral branches of P. alba × P. glandulosa ‘YX2’ (2n = 2x = 38) were collected from a clonal trial in Guan County, Shandong Province (Guan County Nursery). All plant materials were collected from certified nurseries, which guaranteed species identification. All floral branches were collected from 20‑year‑old trees, with a diameter range of 1.5–2.0 cm and a length of approximately 70 cm. The floral branches were cultured in tap water in a greenhouse (10–20 °C) at Beijing Forestry University to force the development of flowers. No additional nutritional components were added to the tap water. Cytology After cultivating the male floral branches, the flower buds were randomly sampled at intervals ranging from 2 to 4 h, and two to three buds were collected. These buds were fixed in Carnoy’s solution until the tetrads emerged. The buds were stored at a 4℃ controlled temperature room for 24 h following fixation, and the anthers were dissected from the buds for a meiotic analysis using forceps and were crushed in a droplet of aceto-carmine solution (2%) on a microscopic slide. Photomicrographs of developing PMCs were taken under a light microscope (model BX61; Olympus, Tokyo, Japan) with a CCD camera (model DP70; Olympus). Approximately 200–300 PMCs were observed in each sample to determine the dominant meiotic stages. The male meiotic products were stained with Aniline Blue solution [0.1% (m/v) in 0.033% K3PO4 (m/v)] following the method reported by Liu et al., [19]. The stained samples were examined and photographed using a Leica TCS-SP8 confocal laser scanning microscope. Tubulin immunolocalization During the initiation of meiosis in the PMCs, the P. tomentosa anthers were sampled at various meiotic stages and subjected to fixation in 4% paraformaldehyde for 45 min. α-Tubulin immunolocalization was conducted following the methods outlined by Zhang and Kang [20]. The slides were washed three times with PBS, each wash lasting 5 minutes, and were counterstained with 20 µl of 4’, 6-diamidino-2-phenylindole. The Leica TCS-SP8 confocal laser scanning microscope was used to examine and photograph the samples. 2n pollen induction by ELT When the dominant meiotic stages of the PMCs were metaphase I, prophase II, metaphase II, anaphase II, and telophase II, the PMCs were subjected to low temperature for 3, 4, and 5 h at − 7, −10, and − 13 ℃, respectively. Untreated male flower buds served as the control. All treated male flower branches were hydroponically cultured until the pollen was released from the anthers. When the anthers matured, five catkins were randomly sampled to determine the frequency of induced 2n pollen. After collecting the pollen, the pollen samples were stored in glass bottles containing allochronic silica gel. The frequency of induced 2n pollen was estimated according to the method reported by Zhang and Kang [20]. The diameters of 300–400 pollen grains were measured in each sample. Triploid production by crossing ELT-induced 2n pollen When the stigmas of female flower buds were receptive, they were pollinated with fresh induced 2n pollen of P. tomentosa clone 145 and P. alba × P. glandulosa ‘84K’. The female floral branches were hydroponically cultured. Seeds matured after approximately 4 weeks of cultivation and were collected and germinated in sterile soil. When the seedlings had grown to approximately 5 cm in height, they were transferred to containers and fertilized to promote growth. Subsequently, surviving seedlings were transplanted in the field at a height of approximately 25 cm. Ploidy analysis by flow cytometry and chromosome counting Flow cytometry was employed to determine the ploidy level of the offspring according to Galbraith et al., [21]. Nuclei were stained with 4’, 6 - diamidino − 2 - phenylindole (10 mg / ml). Chicken red blood cells (2.33 pg / 2 C) were used as an internal standard, and a leaf sample from a known diploid P. tomentosa (2n = 2x = 38) plant was employed as the external standard. The C-value of Populus is 0.46 pg / 1 C (Kew C-value database), and this value corresponds to the standard peak observed on relative fluorescent intensity channel 50. Ploidy analysis by flow cytometry and somatic chromosome counting of progeny were done according to the methods documented by Zhou et al., [22]. The triploid production rate was estimated by the number of real triploids confirmed via chromosome counts and the sum of surviving seedlings in each treatment. Statistical analysis We analyzed the frequency of ELT-induced 2n pollen using the GLM-univariate to reveal the differences between meiotic stages, temperatures, durations, and their interactions. A transformation of the data was done for the frequency of ELT-induced 2n pollen (1/p) before analyses of variance to account for heterogeneity variance. Statistically significant differences among groups were determined using LSD test, and different lowercase letters were used to indicate significant groupings (P < 0.05). All statistical analyses were performed using SPSS software (SPSS for Windows, Version 18.0, SPSS Inc., Chicago, IL, USA). Results Meiosis of P. tomentosa PMCs Meiotic analysis of PMCs was used to guide the doubling of gamete chromosomes to obtain 2n pollen by a particular physical and chemical mutagen. Meiosis of PMCs was consecutive and asynchronous in P. tomentosa clone 145 (Table 1). More than one meiotic stage was seen in each male bud owing to the asynchronous development of PMCs in different anthers, similar to other Populus species. Table 1Meiosis of PMCs in P. tomentosa clone 145Meiotic stage of PMCsHours after culture (h)3244526066727577798081Leptotene79.527.55.0Zygotene17.556.010.03.50.5Pachytene3.012.5276.03.5Diplotene4.03822.513.5Diakinesis13.546.021.5Metaphase I6.519.528.59.02.5Anaphase I2.514.524.57.00.5Telophase I16.538.017.09.53.51.0Prophase II1.521.040.021.59.52.0Metaphase II5.021.541.528.56.01.0Anaphase II2.511.019.539.514.52.5Telophase II1.07.517.046.514.0Tetrad2.030.082.5 Meiosis of PMCs in P. tomentosa clone 145 The PMCs underwent meiosis initially after 32 h of greenhouse culture. Many PMCs were at leptotene (79.5%; Fig. 1A). The dominant meiotic stage varied from zygotene to diakinesis after 44–60 h of culture (Fig. 1B - E). After the male flower buds were cultured for 66 h, most of the PMCs were at metaphase I (28.5%; Fig. 1F). Some PMCs were observed at other stages, including zygotene, pachytene, diplotene, diakinesis, anaphase I (Fig. 1G), telophase I (Fig. 1H), and prophase II, as shown in Fig. 1I. We observed anaphase I (24.5%) after 72 h of culture. Telophase I (37.5%) was detected under light microscopy, indicating that the first cell division was complete. It took approximately 40 h for the PMCs to complete the first meiotic division. Meiosis of PMCs from P. tomentosa clone 145. A Leptotene. B Zygotene. C Pachytene. D Diplotene. E Diakinesis. F Metaphase I. G Anaphase I. H Telophase I (I) Prophase II (J) Metaphase II. K Anaphase II. L-N Telophase II. O Tetrad. Bars = 5 μm The second meiotic division was directly initiated without cytokinesis. Prophase II (40.0%) became the dominant stage after 75 h of culture. Additionally, we observed metaphase I, anaphase I, telophase I, metaphase II (Fig. 1J), and anaphase II (Fig. 1K). The dominant meiotic stage varied from metaphase II (41.5%) to telophase II (46.5%; Fig. 1L - N) during 77–80 h of culture. Tetrads (82.5%; Fig. 1O) were the dominant stage when the cell plates formed after 81 h of culture. The second division was significantly shorter than the first division, only lasting a total of 9 h. Short periods of ELT causes defects in male meiotic cytokinesis of PMCs by interfering with RMA formation at telophase II Microtubular phragmoplast-like structures are necessary for male meiotic cytokinesis. Nevertheless, they are referred to as radial microtubule arrays (RMAs) owing to their specific spatial intercellular organization of forming at the intersection of microtubules (MTs) emanating from syncytial telophase II nuclei [23]. Due to the importance of RMAs in the formation and positioning of the male meiotic cell wall [24], we performed immunocytological analysis of the MT subunit α - tubulin in control and ELT - treated PMCs (Fig. 4). In the control group, PMCs at leptotene exhibited a nucleus comprised of chromatin and a nucleolus, with MTs detected in the perinuclear region (Fig. 2A). Subsequently, the chromatin condensed into chromosomes, accompanied by some MTs entering the nucleus; all MTs were extended throughout the nucleus at diakinesis (Fig. 2B). The chromosomes were aligned at the spindle equator at metaphase I (Fig. 2C), and then the homologs moved to the spindle poles at anaphase I (Fig. 2D). The MTs extended throughout the cytoplasm at telophase I (Fig. 2E). The chromatin immediately condenses into chromosomes again. Progressing to prophase II, the MTs entered the two daughter nuclei (Fig. 2F). The chromosomes aligned at the spindle equator at metaphase II (Fig. 2G). The sister chromatids separated during anaphase II and migrated towards the spindle poles of the respective daughter nuclei (Fig. 2H). Consequently, four daughter nuclei were generated, and the RMAs were extensively distributed throughout the cytoplasm at telophase II (Fig. 2I). Four microspores were released at the tetrad stage following meiotic cytokinesis (Fig. 2J). Tetrads, triads, and dyads in the ELT-treated and control groups. A Tetrads in the control group. B Dyads were produced after exposure to -13 °C for 3 h (arrow). C Dyads were produced after exposure to -13 °C for 4 h (arrow). D Dyads and triads were produced after exposure to -13 °C for 5 h (arrow). Bars = 50 μm Meiotic microtubule (green) and chromosome (blue) arrangements during male meiosis in the control and ELT-treated PMCs of P. tomentosa clone 145. A-J Microtubule and chromosome arrangements during meiosis in control PMCs. A Leptotene in the control group. B Diakinesis in the control group. C Metaphase I in the control group. D Anaphase I in the control group. E Telophase I in the control group. F Prophase II in the control group. G Metaphase II in the control group. H Anaphase II in the control group. I Telophase II in the control group. J Tetrad in the control group. K - O Microtubule and chromosome arrangements during meiosis in the − 13℃ ELT-treated PMCs. K Metaphase II in the − 13℃ ELT-treated PMCs. L Anaphase II in the − 13℃ ELT-treated PMCs. M - O Defects in RMA formation of ELT-treated PMCs at telophase II. Bars = 5 μm After the PMCs were treated with ELT (-13℃ for 5 h) at metaphase II, their MTs were similar to those of the control group at metaphase II (Fig. 2K) and anaphase II (Fig. 2L). However, some distinct RMAs between each of the four haploid nuclei were seen at telophase II (Fig. 2M - O), which were not seen in the control group, indicating that ELT changed the biogenesis and organization of the RMAs. In some cases, the RMAs did not form or form partially in some of the treated PMCs (Fig. 2M - O). This was not observed in the untreated control group, suggesting that ELT altered the biogenesis and organizational patterns of the RMAs. Moreover, some ELT-treated tetrads with regularly separated nuclei exhibited poor inter-nuclear MT labeling, indicating minor aberrations in RMA biogenesis or MT stability (Fig. 2M, N). The accumulation of intermediate callosic material at the end of meiosis II reflects proper formation of the cell wall and cytokinesis during male meiosis in Populus species. Aniline Blue staining of ELT-treated PMCs revealed changes in the deposition of callosic material in the cell wall (Fig. 3). Almost all of the PMCs in the control group generated normal tetrads (Fig. 3A). In some cases, ELT-treated PMCs also produced normal tetrads (Fig. 3B), similar to those observed in the control, and typically characterized by a distinct cross-shaped callosic cell wall separating all four haploid spores. However, a subset of tetrad-stage meiocytes in other ELT-treated PMCs produced partial or incomplete callosic cell walls at the end of meiosis II, resulting in the production of balanced and unbalanced tetrads, triads, and dyads (Fig. 3C - H). A large variation in the severity of the defect was detected in this pool of altered meiotic products; some cells did not form a callosic wall (Fig. 3E, F), whereas others generated cell wall stubs or cell plates with small gaps (Fig. 3C, D). Defects in male meiotic cytokinesis in ELT-treated PMCs of P. tomentosa clone 145. A Aniline Blue staining of the tetrads in the control. B Aniline Blue staining of the tetrads in the ELT treatments. C - H Aniline Blue staining of defective tetrad stage male meiocytes in the ELT treatments. Bars = 5 μm RMA defects during male meiotic cytokinesis in Populus result in triads and dyads. Therefore, we subsequently counted the percentage of meiotic products in the ELT treatments. Only tetrads (Fig. 4A) were observed at the tetrad stage in the control group, and no triads or dyads were observed (Table 2). However, in the ELT treatment (-13 °C for 3 h), 81.61% tetrads, 0.32% triads, and 18.06% dyads (Fig. 4B) were observed; 74.06% tetrads, 1.71% triads and 24.23% dyads (Fig. 4C) were observed in the ELT treatment (-13 °C for 4 h); and 61.58% tetrads, 4.27% triads, and 34.14% dyads (Fig. 4D) were detected in the ELT treatment (-13 °C for 5 h)(Table 2). The percentages of dyads in the ELT treatment were much higher than that in the control. The percentages of triads in the ELT treatment were slightly higher than that of the control. Therefore, we concluded that the abnormal formation of the cell plate led to the formation of dyads and triads. Overall, our data support that meiotic restitution resulting from defective male meiotic cytokinesis caused by defects in RMA formation. Table 2Percentage of meiotic products in the − 13℃ stress at metaphase IIExposure time (h) n Percentage of meiotic productsDyadTriadTetrad333418.060.3281.61431024.231.7174.06529334.144.2761.58Control32800100 Percentage of meiotic products in the − 13℃ stress at metaphase II Short periods of ELT induces P. tomentosa clone 145 to produce 2n pollen Because defects in RMA formation can be caused by ELT, we speculated that ELT treatment would induce male Populus to generate 2n pollen by interfering with male meiotic cytokinesis. To investigate the effect of ELT on male sporogenesis in Populus, male flower buds of P. tomentosa clone 145 were subjected to -7℃, -10℃, and − 13℃ for 3, 4, and 5 h, respectively, within a controlled thermostat environment. Some buds died after the ELT treatment, resulting in no pollen production. Surviving flower buds were continuously cultured in the greenhouse until the anthers matured. The ploidy of the pollen was analyzed after the ELT treatment to assess the putative changes in male spore formation. The ploidy level of the induced pollen grains was determined by measuring the pollen diameter, which is a proxy for the ploidy level of the gametophyte [25]. A few natural 2n pollen grains were observed in the control group (Fig. 5A). The frequency of natural 2n pollen was 0.48%. ELT-induced 2n pollen was collected from all the surviving treated male buds (Fig. 5B - D). The frequencies of induced 2n pollen derived from the different treatments are presented in Table 3. The average frequency of induced 2n pollen varied from 1.26% to 31.11%. A GLM-univariate analyses of the frequencies of ELT-induced 2n pollen demonstrated that the dominant meiotic stage (F = 19.804, P < 0.001), low temperature (F = 30.656, P < 0.001), and duration (F = 5.554, P = 0.005) had significant effects on the frequency of induced 2n pollen. Then, LSD multiple-comparison tests showed that significantly more 2n pollen was induced at metaphase II than at the other stages (P < 0.05). The frequency of induced 2n pollen was significantly higher in samples exposed to -13℃ than in those exposed to -7℃ or -10℃ (P < 0.05). The frequency of induced 2n pollen in samples after the 5 h ELT treatment was significantly higher than those that received the 3–4 h ELT treatment. Therefore, when the dominant meiotic stage of the PMCs was metaphase II, the 5 h -13 ℃ treatment was optimal for inducing 2n pollen in male floral buds via ELT treatment. Natural 2n pollen and induced 2n pollen via ELT stress in P. tomentosa clone 145 and P. alba × P. glandulosa ‘84K’. A Natural 2n pollen in P. tomentosa clone 145 (arrow). B Induced 2n pollen via ELT treatment exposure to -13℃ for 3 h in P. tomentosa clone 145 (arrow). C Induced 2n pollen via ELT treatment exposure to -13℃ for 4 h in P. tomentosa clone 145 (arrow). D Induced 2n pollen via ELT treatment exposure to -13℃ for 5 h in P. tomentosa clone 145 (arrow). E Natural 2n pollen in P. alba × P. glandulosa ‘84K’ (arrow). F Induced 2n pollen via ELT treatment exposure to -13℃ for 3 h in P. alba × P. glandulosa ‘84K’ (arrow). G Induced 2n pollen via ELT treatment exposure to -13℃ for 4 h in P. alba × P. glandulosa ‘84K’ (arrow). H Induced 2n pollen via ELT treatment exposure to -13℃ for 5 h in P. alba × P. glandulosa ‘84K’ (arrow). Bars = 20 μm Table 3Induction of 2n pollen via ELT treatment in P. tomentosa clone 145Dominant meiotic stage of PMCsLow temperature (℃)Duration (h)Frequencies of induced 2n pollen (%)Metaphase I-731.55 ± 0.0441.84 ± 0.8852.00 ± 0.56-1032.07 ± 0.2642.46 ± 0.8452.85 ± 0.46-1332.89 ± 1.4343.27 ± 0.2554.21 ± 2.52Prophase II-731.68 ± 0.2742.17 ± 1.2252.30 ± 0.89-1032.37 ± 0.8442.84 ± 1.1653.27 ± 1.50-1333.70 ± 0.7744.31 ± 1.6556.11 ± 2.25Metaphase II-732.45 ± 1.0144.28 ± 1.3255.10 ± 1.21-1036.88 ± 1.6248.11 ± 1.63510.22 ± 2.31-13314.92 ± 1.74419.86 ± 3.53531.11 ± 2.93Anaphase II-732.13 ± 0.4942.65 ± 1.1152.74 ± 1.01-1032.86 ± 1.2443.16 ± 1.6253.81 ± 2.73-1334.25 ± 2.0445.64 ± 2.3357.40 ± 3.77Telophase II-731.26 ± 0.2741.56 ± 0.1151.86 ± 0.74-1031.99 ± 0.7742.36 ± 1.5552.56 ± 0.81-1332.79 ± 1.5043.00 ± 1.3253.97 ± 1.49Control0.48 ± 0.16 Induction of 2n pollen via ELT treatment in P. tomentosa clone 145 To test the feasibility of inducing 2n pollen in other Populus species using ELT treatment during male meiosis, we stressed male floral buds of P. alba × P. glandulosa ‘84K’ by exposing them to -13℃ for 3, 4, and 5 h, respectively, when most of the PMCs arrived at metaphase II. All stressed male branches continued to be cultured in the greenhouse until the pollen was released from the anthers. Similar to P. tomentosa clone 145, a few natural 2n pollen grains were observed in the P. alba × P. glandulosa ‘84K’ control group (Fig. 5E), and the frequency of natural 2n pollen generated was 0.34%. Among the ELT-treated groups, a large number of ELT-induced 2n pollen grains was seen in P. alba × P. glandulosa ‘84K’ (Fig. 5F-H). The frequency of ELT-induced 2n pollen ranged from 10.45% to 39.17%, which was significantly higher than that of the control group. Triploid production by crossing ELT-induced 2n pollen According to Tables 4 and 12 combinations of crosses with relatively high frequencies of induced 2n pollen between P. tomentosa and P. alba × P. glandulosa were conducted. A total of 6,167 seeds were obtained from the cross combinations with different frequencies of ELT-induced 2n pollen and the control group. About 1,540 seeds were sown and developed into young seedlings (Table 4). According to the flow cytometry peaks, individual seedlings were described as diploid (Fig. 6A) or triploid (Fig. 6B). Finally, 26 putative triploids were detected. The chromosome number in diploids was 38 (Fig. 6C), and all putative triploids were confirmed to be real triploids (2n = 2x = 57, Fig. 6D) by further somatic chromosome counts, which demonstrated that ELT treatment had no significant effect on 2n pollen viability. All triploids were produced from these cross combinations with different ELT-induced 2n pollen. Six triploids came from the cross combination between P. tomentosa clone 121 and P. alba × P. glandulosa ‘84K’. Twenty triploids were derived from the cross combination between P. alba × P. glandulosa ‘YX2’ and P. tomentosa clone 145. No triploids were detected in the control group, indicating that P. tomentosa clone 121 did not generate 2n eggs, and no natural 2n pollen participated in fertilization of the control group. Table 4Triploid production using ELT-induced 2n pollen from P. tomentosa and P. alba × P. glandulosaCross combinationsFrequencies of induced 2n pollen (%)Seed numberSeedlings numberTriploid numberTriploid production rate (%)P. tomentosa clone 121 × (P. alba × P. glandulosa) ‘84K’10.4513004100012.49503370014.02805490016139736730.8239.17295410030.952156434.69(P. alba × P. glandulosa) ‘YX2’ × P. tomentosa clone 1456.8844612310.818.112858711.1510.2293841011.9014.9217116212.5019.8622128414.2931.113610220.000.48 (control)40022400Total61671540261.69 Triploid production using ELT-induced 2n pollen from P. tomentosa and P. alba × P. glandulosa Ploidy level analysis of offspring derived from female gametes of P. alba × P. glandulosa ‘YX2’ crossing ELT-induced 2n pollen in P. tomentosa clone 145 and that from female gametes in P. tomentosa clone 121 crossing ELT-induced 2n pollen in P. alba × P. glandulosa ‘84K’. A Flow cytometry of diploid plants. B Flow cytometry of triploid plants. C Chromosome number of diploid plants (2n = 2x = 38). D Chromosome number of triploid plants (2n = 2x = 57). Bars = 2.0 μm Discussion This study showed the effect of short periods of ELT on pollen formation in P. tomentosa. The formation of RMAs of PMCs was affected by ELT, resulting in meiotic restitution and 2n pollen production. A protocol for inducing 2n pollen in Populus using short periods of ELT was established. The highest frequency of induced 2n pollen (31.11%) was observed when the dominant meiotic stage of the PMCs was metaphase II, and the cells were exposed to -13℃ for 5 h. Twenty-six triploids were produced by crossing induced 2n pollen. Therefore, we developed a reliable and efficient method for triploid breeding of plants. Cytological mechanisms of ELT-induced 2n pollen formation during meiosis Meiosis is a necessary stage of gamete formation during sexual reproduction in higher plants and is a current research interest in reproductive biology. A series of complex processes, such as chromosome duplication, homologous chromosome pairing, homologous chromosome and sister chromatid separation, and cytokinesis, occur during meiosis. Abnormalities at any stage of the process can result in the production of 2n gametes [6]. Irregular chromosome pairing and stickiness of univalent chromosomes in Begonia were reported by Dewitte et al., [26], leading to the production of natural FDR-type 2n pollen. Fused spindles, and premature cytokinesis were also considered to cause dyads in Populus euphratica [20]. In Lotus tenuis and Rosa spp., an increase in parallel spindles induced by high temperatures drove 2n pollen production [12]. Chromosome bridges during meiotic anaphase have also been demonstrated to form 2n gametes [26]. In a recent study, complete loss of meiosis Ⅱ was induced in some PMCs after the PMCs were exposed to 38℃ at diakinesis, resulting in the formation of dyads at the tetrad stage [22]. In our study, after PMCs were exposed to -13℃ during metaphase Ⅱ, no abnormal chromosomal behavior was observed (Fig. 4K-O), indicating that abnormal chromosomal behavior was not responsible for the formation of ELT-induced 2n pollen. The progress of meiosis in higher plants is closely related to the arrangement and changes in the microtubular cytoskeleton. MTs participate in the formation of spindles and the phragmoplast, and microtubule networks are constructed in the perinuclear region and cytoplasm. MTs regulate the separation and movement of homologous chromosomes and sister chromatids [27] and are necessary for cytokinesis [28]. MTs are sensitive to temperature extremes and display rapid depolymerization in response to adverse temperatures [29, 30]. Smertenko et al., [31] reported the structure of the microtubule arrays in tobacco (Nicotiana tabacum) cells incubated at 38℃ or 42℃: the majority of the MTs were disrupted after 30 min at 42℃, whereas all of the MT arrays were preserved after 12 h at 38℃. Storme et al., [15] showed that ELT stress alters the formation of RMAs at telophase Ⅱ and consequently causes defects in post-meiotic cytokinesis and cell wall formation in Arabidopsis thaliana. In this study, after PMCs were exposed to -13℃ at metaphase Ⅱ, the ELT treatment changed the formation of RMAs at telophase Ⅱ, and some defects in post-meiotic cytokinesis and cell formation were observed in P. tomentosa clone 145, similar to the Arabidopsis findings. Many dyads were observed at the tetrad stage, suggesting that short period of ELT induced Populus to produce 2n pollen by destabilizing the post-meiotic RMAs. Short period of ELT treatment and 2n pollen production Here, ELT treatment has been highlighted as a single environmental factor that induces a high frequency of 2n pollen, thus artificially generating triploids. We successfully developed ELT treatment into an efficient mutagen for inducing 2n pollen using a combination of α - tubulin immunolocalization and pollen chromosome doubling technology. In particular, treating Populus male flower buds with ELT (from − 7℃ to -13℃) caused defects in male meiotic cytokinesis by interfering with the formation of RMAs, resulting in meiotic restitution and 2n pollen production. Then, we applied ELT to treat the male flower buds at different meiotic stages, and a large number of ELT-induced 2n pollen grains were produced by P. tomentosa clone 145 and P. alba × P. glandulosa ‘84k’. Twenty-six triploids were produced by crossing induced 2n pollen. Therefore, ELT could be used as an efficient physical mutagen to induce 2n pollen production. Populus PMCs are sensitive to the environmental temperature during meiosis, and the range of efficient temperatures is narrow. Precise temperature is necessary to induce 2n pollen production in plants. For example, Pécrix et al., [12] reported that 2n pollen production after exposure to 36 °C for 48 h was significantly higher than that of the control, and the 2n pollen production rate was 24.5% in Rosa spp. The precise high temperature to induce 2n pollen production in P. tomentosa × P. bolleana was reported to be 38 °C and the 2n pollen production rate was more than 80%. A temperature of 38 °C was also recorded as the precise high temperature to induce 2n pollen production in P. canescens [22]. The highest frequency of induced 2n pollen (32.75%) was obtained in the 36 °C 4 h treatment of P. alba [32]. Mao et al., [33] reported that the 2n pollen production rate was maximal after a 4 h 46 °C treatment of Eucommia ulmoides. However, in the present study, the precise ELT to induce 2n pollen in P. tomentosa was − 13 °C and the highest frequency of induced 2n pollen was 31.11%. Excessively low temperatures can cause pollen abortion and even flower bud death. By jointly considering the dual effects of ELT on male flower bud survival and pollen viability, we determined that the optimal temperature range for inducing 2n pollen in P. tomentosa is -13 °C to -7 °C. This range balanced effective induction with the maintenance of tissue health and acceptable pollen yield. The precise temperature varied depending on the species or genotype. Therefore, the precise temperature should be adjusted according to different species or genotypes when 2n pollen production is to be induced. Many studies have reported that the sensitivity window is limited to meiosis [12], and not all meiotic stages are available for 2n pollen induction by changing the environmental temperature. Therefore, it is more challenging to determine the accurate meiotic stage during asynchronous meiosis of PMCs in Populus. In the present study, the developmental stage with the highest frequency was defined as the dominant meiotic stage. The induction of 2n pollen via ELT was performed at a particular dominant meiotic stage. Applying a mutagenic agent to cells that are at a suitable meiotic stage is important for inducing chromosome doubling during gametogenesis [34]. Previous studies have shown that the most effective stage for inducing 2n pollen with colchicine is pachytene [35, 36]. Tian et al., [37] reported that diakinesis is the most suitable meiotic stage for inducing 2n pollen via high temperature in P. canescens. Nevertheless, when the PMCs of P. tomatosa at metaphase II were stressed at -13 °C, the highest frequency of ELT-induced 2n pollen (31.11%) was obtained, suggesting that the most effective stage for inducing 2n pollen by ELT was metaphase II. The genetic composition of induced 2n gametes depends on the formation mechanism. In this study, the formation of P. tomentosa PMC RMAs was affected by the ELT, resulting in defects in male cytokinesis. However, how the four daughter cells fuse and form into dyads at the tetrad stage after meiosis of ELT-stressed PMCs remains unknown. Therefore, further work should focus on the cytological and molecular mechanisms of ELT-induced 2n pollen formation in Populus. Conclusions In this study, we found that short periods of ELT could cause defects in male post-meiotic cytokinesis by interfering with the formation of the radial microtubule array, resulting in meiotic restitution and 2n pollen production in Populus. A protocol for inducing 2n pollen by ELT was also established. The dominant meiotic stage, low temperature, and duration significantly affected the frequency of induced 2n pollen. The highest frequency of induced 2n pollen (31.11%) was seen when the dominant meiotic stage of the pollen mother cells was metaphase II and was stressed by -13℃ for 5 h. Twenty-six triploids were generated by crossing induced 2n pollen. Our findings provide new insight into the mechanisms of induced 2n pollen formation in a woody plant and verify that sensitivity to environmental stress has evolutionary importance in terms of polyploidization. Acknowledgments We thank Meiqin Liu for her scanning electron microscopy analysis. Statement of plant collection The researches on cultivated plants comply with relevant institutional, national, and international guidelines and legislation. Authors’ contributions PZ designed the research. ML, YX, DC, XM, DC, BK, YZ and YS performed the research. JD analyzed the data and wrote the original draft. PZ reviewed and edited the paper. All authors contributed to the article and approved the submitted version. Funding This work was supported by the National Key R&D Program of China during the 14th Five-year Plan Period (2022YFD2200301-02). Data availability The data supporting the findings of this study are available within this manuscript. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The data supporting the findings of this study are available within this manuscript.