Introduction

Dioecy (the presence of distinct male and female plants) has evolved repeatedly from cosexual ancestors in nearly half of the angiosperm families (Geber et al. 1999). Dioecious plants exhibit sexual dimorphism in both reproductive organs and vegetative traits (Liu et al. 2021). Selection for outcrossing and compensatory resource reallocation following the loss of one sexual function are thought to be the principal mechanisms favoring sex dimorphism (Miller and Venable 2000). Resource allocation commonly differs between males and females, and these differences are assumed to be associated with trade-offs between reproductive cost and other functions (Seethapathy et al. 2021). For example, many studies have shown that growth performance differs between males and females (reviewed by Lloyd and Webb 1977).

Populus deltoides (cottonwood poplar) is native to North America and an important fast growing tree that can grow up to 40 m (Kartesz and Meacham 1999). Due to its extraordinary growth rate, P. deltoides and its hybrids have been widely used as a commercial source of fiber and fuel in many regions of the world, including North America, Europe, China, India, New Zealand and Australia (Zsuffa et al. 1996; Jain and Singh 2000; Broeck 2005; Wan et al. 2015). P. deltoides is fully dioecious; the androecia are produced on different individuals than the gynoecia. Early empirical observations revealed male superiority in vigor and form in natural populations of poplars (Farmer 1964; Grant and Mitton 1979); thus, the sex of a poplar tree may play an important role in its growth performance. In addition, female poplars produce abundant seed hairs when they reach sexual maturity. A mature female tree bears a huge number of female catkins and produces more than a kilogram of seed hairs. Seed hairs disperse in May of each year through mid-August, causing severe airborne pollution. For the above reasons, it is more desirable to breed elite male cultivars. However, it takes 6–7 years for P. deltoides to reach sexual maturity, so a method to discriminate the sex of P. deltoides at the seedling stage is in demand.

Sex in P. deltoides is governed by the XY determination system. The proximal end of the Y-chromosome harbors a male-specific sequence containing two genes, FERR-R and MSL. FERR-R is the female repressor and MSL is found to function as the male promotor. Expression of FERR-R and MSL triggers the differentiation of the opposite sexes of floral primordia (Xue et al. 2020). This research advance enables us to develop a marker-aided technique to discriminate between sexes in juvenile P. deltoides.

Using the trees in our large germplasm collection of P. deltoides that have been grown in the field for 20 years, we analyzed the effect of sex on growth performance of P. deltoides and developed a marker-aided technique to discriminate the sex of seedlings of P. deltoides to provide essential information and accelerate improvement programs.

Materials and methods

Field trial

A total of 270 P. deltoides clones were collected from natural populations in the United States: 29 from Texas, 95 from Louisiana, and 146 from Mississippi. These P. deltoides clones were maintained at Sihong Forest Farmer, Jiangsu Province, China. The location lies in a transition zone of north subtropical to warm temperate zone, with yellow fluvo-aquic soil and mean annual temperature of 14.6 °C and precipitation of 893.9 mm. Four ramets for each clone were planted in the field using a randomized block design. After 20 years, some of the stands had died, and 914 stands remained in this field trial.

Phenotypic data collection

We recorded sexes of these P. deltoides trees based their floral characteristics; male flower buds expand faster than the female ones, yielding a larger bud. When the flower buds are fully flushed, the most distinctive feature is that the male inflorescences bear florets only with stamens, and females bear florets only with pistils (Fig. 1). At this stage, the sex of the tree can be clearly discerned. We also measured the diameter at breast height (DBH) of each tree. To obtain the basic wood density (BWD), we took a tree core at a stem height of 1.3 m from each genotype. The cores were cut into pieces < 6 cm long, and three pieces per core were used to measure the BWD using the method and formula described by Olesen (1971).

Fig. 1
figure 1

Male and female inflorescence of P. deltoides

Full size image

Marker-aided sex discrimination

In an early study (Xue et al. 2020), we discovered the Y-specific hemizygous sequence (~ 34.8 kb) in the sex determination region on the telomeric end of chromosome 19 in the genome of P. deltoides. In our 2020 study, we re-sequenced the genome of 115 P. deltoides trees (68 females and 47 males) and determined a conserved Y-specific sequence that harbored only two genes: FERR-R (~ 16 kb) and MSL (~ 3 kb). Based on the Y-specific hemizygous sequence, primers were designed and synthesized to amplify DNA from three male and three female trees in a PCR as described by Yin et al. (2008). Primers that generated a distinct amplicon in the males but no amplicon in the females were further verified using natural stands of P. deltoides.

Statistical analyses

We used Microsoft (Redmond, WA, USA) Excel to calculate the basic statistics for the phenotypic data. The sex ratio was analyzed using a χ2 goodness of fit test. Fisher's least significant difference (LSD) test was used to compare the DBH or BWD significant differences among different provenances. Phenotypes were tested for significant differences between sexes using independent-sample t-test in SPSS 19.0 (Coakes and Steed 2003).

Results

Sex ratio of P. deltoides germplasm based on morphology

Among the 270 clones, 141 were discriminated as females, 129 as males; thus, females were more abundant in this P. deltoides collection. However, the χ2 test revealed that the sex ratio did not significantly departure from 1:1 (χ2 = 0.5333 < χ20.05 = 3.841). Over the 20 years of growth, 66 trees had died; 32 were females, and 34 were males, indicating that mortality was not associated with sex.

Variation in DBH and sex effect on DBH

The DBH frequency distribution of each group had a bell-shaped form (Fig. S1), and statistics for radial growth for these P. deltoides are listed in Table 1. The mean DBH was 31.45 ± 4.14 cm and varied from 20.8 to 45.9 cm, with a CV of 13.17%. The maximum was > twofold larger than the minimum, indicating great breeding potential for radial growth. Clones from Texas had higher absolute radial growth than those from Louisiana or Mississippi. However, an analysis of variance revealed no significant difference in radial growth among the three provenances (P > 0.05).

Table 1 Statistics of radical growth of the P. deltoides collections
Full size table

For the males, the mean DBH was 32.28 ± 3.74 cm, varied from 22.2 to 43.3 cm with CV of 11.58%; mean DBH of females was 30.69 ± 4.36 cm (20.8–44.9 cm), with CV of 14.19%. An independent-sample t-test indicated a significant difference in radial growth between the two sexes (P = 0.002). Therefore, our results clearly indicated that gender had a significant effect on radial growth of P. deltoides.

Variation in BWD and sex effect on BWD

The shape of the BWD frequency distribution of each group was represented by a unimodal curve (Fig. S2). Statistics for BWD for this P. deltoides collection are listed in Table 2. Our measurements showed that the BWD values were more stable than that of DBH, with a CV of 5.61% vs. 13.17%. The mean BWD for these P. deltoides was 0.38 ± 0.02 g cm−3, varied from 0.34 to 0.47 g cm−3. The highest BWD was 138% greater than the lowest. Thus, there is a considerable variance for BWD in the natural P. deltoides population. Correlation analysis revealed a weakly negative correlation between BWD and DBH (correlation coefficient = − 0.143, significant at P < 0.05). For the three provenances, there was little difference in mean BWD values, and LSD analysis showed no significant differences for BWD among the three provenances (P > 0.05).

Table 2 Statistics of density of wood for the P. deltoides collections
Full size table

For the males, the mean BWD was 0.39 ± 0.02 g cm−3 (0.33–0.46 g cm−3) with CV of 5.75%; females had the same mean BWD as the males (0.34–0.47 g cm−3) with CV of 5.49%. The independent-sample t-test revealed that sex had no significant effect (P = 0.569) on BWD.

Marker-aided discrimination of the alternate sexes

The PCR using the primer pairs based on the conserved Y-specific hemizygous sequence of P. deltoides (Xue et al. 2020) yielded the expected sizes for the PCR products (200–300 bp). Using DNA from three male and three female trees, we further screened primer pair N1869 (F: GGACAACACACAAGGCTCTGC; R: TCGTGGGTTGATGTGTAGAGG), N1871 (F: TGCAAGTACAACAGGAGTTCAA; R: TCCACCGACAAGAGACAAAGA), and N1875 (F: CAATATTTGTATTGCCCGTAG; R: CCACTGTTCCATGTTTTAAGG), which generated distinct bands of the expected sizes. Based on their positional information, N1871 and N1875 are located within the genic region of MSL, and N1869 is in the intergenic region between MSL and FERR-R. Since the hemizygous sequence is Y-specific, the expected PCR products were only generated in the males. Thus, the sex of P. deltoides was determined by the PCR products generated, and the three primer pairs provide a triple confirmation of the result.

We further verified the reliability of this marker-aided sex discrimination technique using 36 male and 36 female P. deltoides trees for which the genome had been re-sequenced. Results showed that all males had the expected bands and the females had none (Fig. 2). The technique thus had 100% accuracy for discriminating the sex of trees of P. deltoides and will be a reliable marker-aided tool to determine the sex of very young seedlings and accelerate breeding programs for P. deltoides.

Fig. 2
figure 2

Polyacrylamide gel electrophoresis of PCR products from the alternate sexes of P. deltoides trees. Note a corresponds with N1869; b corresponds with N1870; and c corresponds with N1875. The upper panel (lanes 1 − 36) shows the males and the lower panel shows the females in each chart. M represents the molecular weight standard

Full size image

Discussion

Numerous studies have shown sex-based differences in secondary traits, including morphological, physiological and ecological characteristics (e.g., Dawson and Bliss 1989; Laporte and Delph 1996; Espirito-Santo et al. 2003), and the deployment of sexual function into different morphs may be modified by environmental stressors (Liu et al. 2021). Many studies have pointed to the possibility that males perform better than the females in terms of growth, vigor and form, but many others have reported that females performed similarly or even better than the males (e.g., Grant and Mitton 1979; Popp and Reinartz 1988; Hou et al. 2017). Therefore, there is no regular pattern of sex-based differences in performance among dioecious plants. This inconsistency might be related to biological and genetic differences or to experimental error. For tree species, the age and environmental habitat may vary greatly for the natural stands. To investigate a sex effect, it is essential to carry out a common garden experiment, but many studies have used only a limited number of trees, and conclusions were based on responses of juvenile trees due to the large size of mature trees. Here, many replicate trees of P. deltoides had been planted in a common garden spaced at 6 m × 6 m. These 20-year-old trees provided desirable plant materials in a large-scale field trial to investigate sex effects on growth performance of P. deltoides.

Conflicting sex effects have also been documented for species in Salicaceae. The expected sex ratio is 1:1. However, this ratio can be altered by stressors such as water supply and nutrient availability (Aber Krischik and Denno 1990; Costich 1995). In our collections of P. deltoides, the sex ratio did not significantly depart from 1:1, in accordance with early empirical observations of natural populations of P. deltoides (Farmer 1964) and P. tremuloides (Einspahr 1960). It was noteworthy that our study revealed male superiority in radial growth of P. deltoides as reported by Farmer (1964), but no sexual difference in growth performance was observed in P. tremuloides (Einspahr 1960). For Salix rigida, males were reported to grow faster than the females (Elmqvist et al. 1991), but females of some other willow species grew as fast as or faster than males (e.g., Crawford and Balfour 1990; Åhman 1997; Hou et al. 2017). Thus, sex-based effects may vary among species and among the traits investigated. As noted already, P. deltoides has the XY sex determination system (Xue et al. 2020). S. suchowensis (Hou et al. 2015), S. viminalis (Pucholt et al. 2015) and S. triandra (Li et al. 2020) all have the ZW sex determination system. Males of P. deltoides were shown to have superior growth in our study and by Farmer (1964), whereas females were superior for a variety of traits in the three ZW willow species mentioned above (Åhman 1997; Hou et al. 2017; Yang et al. 2020). Since the sex determination system of plant species has been unclear in most previous studies, whether sex-based differences are associated with the sex determination systems among dioecious plants remains unknown.

For breeding commercial trees, growth rate and wood quality are among the most important traits. The present study showed that great variations in DBH and BWD in the natural stands of P. deltoides, indicating a great potential for improvement. In a study of numerous poplar hybrids, faster growth was generally related to lower wood density (Zhu et al. 1993), but no significant correlation between growth rate and wood density was found for S. suchowensis (Hou et al. 2017). Here, we found a weak negative correlation between DBH and BWD, indicating it is possible to breed elite P. deltoides cultivars with fast growth rate and good wood quality. Wood density is a fundamental wood quality trait. To date, studies of sex effects on wood quality have been scarce. However, our study showed no sex-based difference in wood density of P. deltoides, consistent with the findings for S. suchowensis (Hou et al. 2017).

Of the three sex-discriminating markers based on the Y-specific hemizygous sequence of P. deltoides, two were located within the genic region of MSL. Both the ZZ/ZW and XX/XY sex-determining systems have been found in Populus species. Although a Y-specific hemizygous region is also present in other Populus species that have an XX/XY system, such as P. davidiana (Xue et al. 2020), P. tremula (Pakull et al. 2015) and P. trichocarpa (Müller et al. 2020), the Y-specific hemizygous sequence vares greatly in different poplar species. Therefore, a specific sex discrimination toolkit needs to be developed for different Populus species; the sex-discriminating toolkit developed in this study can be only used for P. deltoides.

In conclusion, male trees of P. deltoides were superior to females in radial growth; thus, considering with seed-hair pollution from the females, male sex should be a selection focus in breeding programs of this dioecious plant. We developed a reliable technique to discriminate the two sexes in seedlings of P. deltoides, which will accelerate breeding programs to improve this important commercial tree.