Introduction

Triticale (× Triticosecale Wittm.) is a man-made cereal developed by crossing wheat (Triticum aestivum L.) and rye (Secale cereale L.) with a genomic constitution of 2n = 6 ×  = 42 (Ayalew et al. 2018; Mergoum et al. 2019). It combines valuable traits like high fertility and grain quality received from wheat with high resistance to diseases, drought, low temperature, higher tolerance to environmental stress factors, and winter hardiness obtained from rye (Tyrka and Chełkowski 2004; Pronyk and Mazza 2011; Mergoum et al. 2019; Golebiowska-Paluch and Dyda 2023). All those qualities make triticale a valuable and well-established worldwide crop which grain is used mainly for animal nutrition but also for human consumption (Losert et al. 2017; Ayalew et al. 2018).

Triticale is also a valuable genetic bridge for transferring eligible genes among rye and wheat genetic pool using molecular breeding techniques (Tams et al. 2004; Tyrka and Chełkowski 2004; Badea et al. 2011). Quantitative Trait Loci (QTL) mapping, Molecular-Assisted Selection (MAS) technique, Genomic Selection (GS), and Next-Generation Sequencing (NGS) are currently widely used in improving crop species and in modern breeding programmes, including wheat and rye (Ahmed et al. 2021; Hackauf et al. 2022; Tsenov et al. 2022). In contrast, molecular breeding in triticale is still limited (Jannink et al. 2010; Tyrka et al. 2018; Mergoum et al. 2019; Dyda et al. 2022a, b; Golebiowska-Paluch and Dyda 2023). So far, a few research studies describe genetic maps of triticale (Alheit et al. 2011; Tyrka et al. 2011, 2015, 2018; Dyda et al. 2022a; Wąsek et al. 2022) together with QTL and candidate genes identification for important traits in this cereal specie like resistance to Xanthomonas vasicola, which causes bacterial leaf streak (Wen et al. 2018), resistance to powdery mildew caused by Blumeria graminis at the adult plant stage (Karbarz et al. 2020; Dyda et al. 2022a,b), cold acclimation and Microdochium nivale tolerance/susceptibility (Dyda et al. 2019; Gołębiowska et al. 2021), and freezing tolerance (Wąsek et al. 2022). The studies on the genetic control of morphological and yield-related traits in cereals have been mainly focused on adult plants of wheat, barley, maize, and rice (Peng et al. 2011; Li et al. 2016, 2020; Gao et al. 2016; Kumar et al. 2016; Mora et al. 2016; Kumari et al. 2018; Cao et al. 2019; Khahani et al. 2020; Kuang et al. 2020; Xin et al. 2020; Hackauf et al. 2022; Tsenov et al. 2022). In contrast, not many triticale studies have been reported so far (Wajdzik et al. 2019; Neuweiler et al. 2020; Karbarz et al. 2020; Gołębiowska et al. 2021; Wąsek et al. 2022). According to our knowledge, there is no research describing the QTLs and candidate genes of any cereal seedling, which is the reason why we present in this paper results that can bring a new insight to understand the genetic control of the seedlings development.

Plant development is a process controlled by genetic information together with environment conditions, e.g. light, soil, temperature, and water (Von Arnim and Deng 1996; Hochholdinger et al. 2004; Chen et al. 2008). Development of cereal seedling is a process controlled by metabolite and hormonal signals, which participate in gene expression, growth, and metabolism; also, sugar regulation of the gene expression at the seedling stage may initiate a specific mechanism, which can participate in the control of this process (Thomas and Rodriguez 1994). Seedling development is a crucial phase in cereal life cycle as it may affect the vitality of adult plant as well as size and quality of the future grain (Stratonovitch and Semenov 2015; Maulana et al. 2018). In winter cereals, the low-temperature period during seedling development (hardening) provides to increasing crop’s low-temperature and fungal pathogens tolerance (Tronsmo et al. 2001; Hudec and Bokor 2002). In our previous studies on triticale, particularly ‘Hewo’ × ‘Magnat’ Doubled Haploids (DH) lines population plant’s hardening led to increased tolerance to Microdochium nivale infection (Gołębiowska and Wędzony 2009; Gołębiowska et al. 2011, 2021; Szechyńska-Hebda et al. 2011, 2013; Dyda et al. 2019; Golebiowska-Paluch and Dyda 2023).

The aim of this study was to (1) locate QTL regions associated with morphological traits of winter triticale tested at a seedling stage and demonstrate the most significant loci, strongly associated with selected features–the width and length of the first and second leaves, the leaf blade and leaf sheath lengths, as well as the length to width ratio of the second leaf, and (2) identify candidate genes/proteins in these regions that could take part in plant development processes.

Materials and Methods

Plant Material and Genetic Linkage Map

The experiments were performed on a population of 89 doubled haploid (DH) lines derived from F1 generation of a cross between two winter triticale cultivars ‘Hewo’ (Strzelce Plant Breeding – IHAR Group Ltd., Poland) as the female parent and ‘Magnat’ (DANKO Plant Breeders Ltd., Poland) as the pollen donor. The population was developed by the anther culture; detailed description of this method and modified protocol was published by Wędzony (2003). Parental cultivars and their offspring differed in morphology at the adult plant stage (Wajdzik et al. 2019), tolerance to Microdochium nivale infection (Gołębiowska and Wędzony 2009; Dyda et al. 2019; Gołębiowska-Pikania et al. 2019; Gołębiowska et al. 2021), freezing tolerance (Wąsek et al. 2022), drought tolerance (Hura et al. 2017), as well as powdery mildew tolerance (Dyda et al. 2022b).

For the above mapping population, the genetic triticale map with total of 4997.4 cM length was developed (Tyrka et al. 2015). All details associated with the method, genotyping, and markers distribution were published by Tyrka et al. (2015). This map was composed of 3539 molecular markers ordered to 20 linkage groups that correspond to triticale genomes A (7), B (7), and R (6). Plant material for all experiments was available at the Institute of Plant Physiology Polish Academy of Science (IPP PAS) in Kraków, Poland.

Plant Growth Conditions and Experiment Design

Analysis of seedlings morphological traits was performed in two-year period in controlled conditions in a completely randomized blocks design (CRBD). Analyses were divided in three independent experiments. Experiments 1 and 2 (Exp. 1, Exp. 2) were performed on cold-hardened 7-week-old seedlings, whereas experiment 3 (Exp. 3) was performed on 4-week-old non-hardened seedlings. Plant treatment with cold (hardening) that was implemented during experiment 1 and 2, was designed to recreate a similar condition for the winter triticale seedling development in natural field conditions where plants are exposed to low temperature. In all three experiments, plants were cultivated as per Gołębiowska and Wędzony (2009) methodology with a detailed modification. All DH lines together with both parental cultivars, were grown in multi-pots in an isolated chamber with 67% humidity. Three series (replicates) per experiment were performed in a CRBD to limit the error resulting from the marginal position of each genotype. 18 kernels of each DH line were sown in 10 rows each, six kernels per genotype/box. One row per each parental line was sown in variable positions in each box as a control (6 kernels per row, 18 kernels per experiment).

In Experiments 1 (1st year) and 2 (2nd year), plants were grown in a temperature regime of 21 °C/16 °C, at 8 h/16 h (day/night) photoperiod for one week and were watered and supplemented with Hoagland and Arnon’s (1938) sterile medium. After that period, seedlings were pre-hardened in a temperature regime of 12 °C/12 °C, at 8 h/16 h (day/night) photoperiod for the next two weeks and subsequently hardened in 4 °C/4 °C, at 8 h/16 h (day/night) photoperiod for the next four weeks. In Experiment 3 (2nd year), to imitate non hardening conditions, plants were continuously grown in 21 °C/16 °C, at 8 h/16 h (day/night) photoperiod in a greenhouse for four weeks, there were watered and supplemented with Hoagland and Arnon’s (1938) sterile medium.

Seedlings Morphology Phenotyping

Seedlings phenotyping was performed on 7-week-old cold-hardened seedlings (Exp. 1 and 2) as well as 4-week-old unhardened seedlings (Exp. 3). It included (1) total number of seedling leaves (NL); (2) lengths of the first (LFL) and second (LSL) leaf measured from the soil surface together with leaf sheath; (3) length of the second leaf sheath from the soil surface to the top of the leaf (LSLS); (4) length of second leaf blade (LSLB) from the base to the top; and (5) widths of the first (WFL) and second (WSL) leaf in the widest point. The unit of length and width was 1 cm. In addition, LSL/WSL (the second leaf length to width) ratio was also measured as a parameter created for the purpose of this research.

Statistical Analysis, QTL Mapping, and Candidate Genes Identification

All raw data (18 seedlings × 91 genotypes × 3 experiments) were analysed using Statistica 13.0 PL software (Statsoft, Tulsa, OK, USA). Distribution of the data was checked using histograms and analysed with a Shapiro–Wilk test, together with skewness and kurtosis. In addition, one-dimensional variance analysis was performed at P < 0.05. Linear correlation coefficients (Pearson’s) were calculated for each of three experiments separately based on mean values that were collected of each replicate per experiment. The regression line was presented with a 95% coefficient interval.

QTL regions were identified based on the mean values for each experiment separately with R/qtl package according to Broman (2012) using composite interval mapping (CIM). The permutations of 1000 times tests were used to determine LOD thresholds to declare a significant QTL (P < 0.05). Significant QTL regions were selected based on the LOD threshold ≥ 3.0. The percentage of the phenotypic variation covered by each QTL was calculated with a single regression factor (R2). Favourable alleles in each QTL region were selected based on the additive effect (Add): negative additive effect referred to cv. ‘Magnat’; positive additive effect referred to cv. ‘Hewo.’ Results of QTL analysis were visualized using CorelDRAW9 software, and the label of each QTL was created from the short name of the morphological trait, chromosome name, and QTL number on the chromosome (1–2).

The candidate genes associated with analysed traits were in silico localized on all identified QTL regions according to method described in detail by Gołębiowska et al. (2021). For this purpose, DArT and e-SNP sequences of the flanking and maximal LOD peak markers of the significant QTLs were used to query all available wheat and rye genome collections for the physical mapping using the BLAST tool of GrainGenes Blast Service beta (https://doi.org/10.1093/molbev/msz185). Genes localized on target physical wheat and rye regions were retrieved and annotated with the use of BLAST® (https://blast.ncbi.nlm.nih.gov/ Blast.cgi). The sequences producing significant alignments and the highest query cover were selected. Next, the function of candidate genes was deduced from the UniProt database by using Quick Go annotation as well as STRING functional network analysis.

Proteomic Analysis

LC–MS analysis was performed exactly as described in detail in previous work (Gołębiowska et al. 2022) on selected seedlings most varied in leaf morphology. Briefly, leaf proteins were extracted, quantified, and digested, and then the recovered peptides were analysed using DionexUltiMateTM 3000 RSLCnano System (Thermo Fisher Scientific, Dreieich, Germany) coupled to an Impact II (Bruker Daltonics, Bremen, Germany). 700 ng of the digested proteins were separated using a Nano Trap Column (Acclaim PepMap100 C18, 5 μm, 100 Å) and 50 cm × 75 µm analytical column (Acclaim PepMap RSLC C18, Thermo Fisher Scientific) through a 120-min-long linear 2–40% acetonitrile gradient (buffer A: 0.1% FA; buffer B: 0.1% FA in ACN) applying a flow rate of 300 nl/min. The CaptiveSpray ion source with a nanoBooster was used to connect the LC system to the MS device. The source was operated in positive ion mode at 150 °C dry temperature, 1300 V capillary voltage, 0.2 bar nanoBooster, and a dry gas flow of 0.3 l/min. For the MS and MS/MS acquisition, the predefined ‘Instant Expertise’ method was used (Compass 1.9, Bruker). The m/z data were acquired in the range of 150 to 2200, and the fixed total cycle time was set to 3.0 s. The instrument settings were as follows: hexapole radio frequency (RF) voltage of 350 V peak-to-peak (Vpp), a funnel 1 RF of 400 Vpp, a funnel 2 RF of 600 Vpp, a prepulse storage time of 10 μs, a transfer time of 90 μs, and a collision cell RF of 2000 Vpp. For the MS spectra, the acquisition speed was 2 Hz with the collision energy of 7 eV. For the MS/MS, the acquisition speed was dependent from the precursor signal intensities and was set to 4 Hz for lower (2500 cts) and 16 Hz for higher (25,000 cts) intensities with linear adjustment for the precursors between low and high. The collision energy was adjusted between 23 and 65 eV as a function of the m/z value. The instrument was calibrated using 10 mM sodium formate. The acquired spectra were analysed for label-free quantifications using Progenesis QI software for proteomics (Version 3.0, Nonlinear Dynamics, Newcastle upon Tyne, U.K.) as recommended by the manufacturer, thereby enabling mass correction, alignment, normalization, peak picking, quantification, and statistics. MS/MS spectra were exported from the Progenesis QI software as Mascot generic files and used for peptide identification with Mascot v2.5.1. The triticale database was created from the rye and wheat sequences (164,836 entries) and merged with the sequences of human keratin and trypsin. Following search parameters were applied: 15 ppm peptide mass tolerance, 0.2 Da fragment mass tolerance, one missed cleavage allowed, carbamidomethylation as fixed modification, and methionine oxidation as variable modification. A Mascot integrated peptide decoy database search was performed, and searches were processed with the Percolator machine-learning algorithm. The false discovery rate was < 1%, ion score cut-off 15, and significance threshold P < 0.05. For subsequent analysis, the set of identified sequences was re-imported into Progenesis QI. For the quantification, a Relative Quantification method with non-conflicting peptides was used. All unique peptides were included; the presence of at least two unique peptides, together with P value below 0.05 and a threshold of 1.5-fold was set as the determinant of significant differential abundance.

Results

Differences in Seedling Leaf Morphology

The cv. ‘Hewo’ seedlings have an upright habit, while cv. ‘Magnat’ seedlings shows lying leaves spread close to the soil surface (Fig. S1). In addition, the dimensions of the leaves differed between parental cultivars in all experiments (Fig. S2). The second leaf sheath and the lamina were on average 0.79 cm and 5.5 cm longer in ‘Hewo’ than in ‘Magnat’ (P < 0.05) 7-week-old cold-hardened seedlings. The second leaf blade was almost 1/3 narrower in ‘Hewo’ than in ‘Magnat’ (P < 0.05) seedlings (Table 1, Fig. S2).

Table 1 Value range of all measured seedling traits of 89 lines of DH ‘Hewo’ × ‘Magnat’ mapping population: number of leaves (NL), the first (LFL) and second leaf lengths (LSL), the first (EFL) and second leaf widths (WSL), the second leaf sheath length (LSLS), and the second leaf blade length (LSLB) as well as LSL/WSL ratio
Full size table

Among the DH lines mapping population, the longest first leaf (LFL) was observed in Exp. 3 (13.3 cm) whereas the longest second leaf (LSL) in Exp. 1 in cold-hardened plants was 25.2 cm. The longest second leaf sheath (LSLS)–4.2 cm–was observed in non-hardened plants in Exp. 3; for plants in Exp. 1 and 2, the length was similar (3.6 and 3.5 cm, accordingly). The longest second leaf blade (LSLB) was observed in Exp. 1 (21.6 cm) comparing to blades in Exp. 2 and 3 (18.8 and 17.4 cm). The first leaf width (WFL) means were the same in Exp. 1 and Exp. 2 (4.1 cm) and, in Exp. 3, the mean observed width value was very similar (4.3 cm). The second leaf width (WSL) means were similar in Exp. 2 and 3 (4.9 cm and 4.6 cm, accordingly) and higher in Exp. 1 (5.7 cm). In addition, the highest number of leaves (eight) was observed in hardened seedlings in Exp. 1, whereas the lowest in Exp. 3 in non-hardened ones (Table 1, Fig. S2).

QTL Identification

All independent factors (plant genotype and experiment number) as well as the interaction between them (plant genotype × experiment number) had significant (P < 0.05) influence on the studied traits (Fig. S2). QTL regions were calculated based on the mean data obtained from individual values from each DH line from each replicate per experiment separately. The Shapiro–Wilk test as well as skewness and kurtosis results confirmed the normal distribution of values for every experiment and each trait (Table 1), what allowed to perform further QTL analysis.

QTL analysis revealed loci associated with all morphological traits measured. However, some of them were statistically insignificant (P > 0.05, LOD < 3.0); thus, the selected described effects, chosen based on the significance, are related exclusively to the second leaf morphology. They include second leaf length (LSL) and width (WSL), second leaf sheath length (LSLS), and blade length (LSLB) together with LSL/WSL ratio (Table 2). Among those results, we have selected the group of effects that contain loci with the highest phenotypic variation (on chromosome 4R) and loci overlapping the same or different traits that appear in more than one experiment (5R and 6R).

Table 2 Characteristics of the most significant QTL regions associated with the second leaf width (WSL), the second leaf length (LSL), the second leaf sheath length (LSLS), the second leaf blade length (LSLB), as well as the second leaf length to width ratio (LSL/WSL) in DH ‘Hewo’ (H) × ‘Magnat’ (M) line seedlings of the winter triticale mapping population
Full size table

Four loci with LOD value ≥ 3.0 associated with the second leaf width (WSL), the second leaf length (LSL), the second leaf sheath length (LSLS), the second leaf blade length (LSLB) and the second leaf length to width ratio (LSL/WSL) were the most significant (Table 2, Fig. 1). Three of those loci appeared in at least two experiments and were mutual for more than one of the analysed traits. Those QTL were associated with the second leaf length and width and were identified exclusively on rye chromosomes 4R, 5R, and 6R with LOD value up to 20.35 (Table 2). No interaction between different QTL was observed. Two common QTL regions were identified on rye chromosome 5R (Table 2). The most significant locus QTL_LS_5R.1 was found on this chromosome in all experiments for LSL, LSLS, and LSL/WSL ratio as well as in one experiment (Exp. 1) for LSLB, with LOD values 3.02–22.35. Positive allele effect of this locus referred to cv. 'Hewo' (Table 2).

Fig. 1
figure 1

Interval map (cM) for chromosomes 4R, 5R, and 6R of the DH ‘Hewo’ × ‘Magnat’ lines mapping population of winter triticale (× Triticosecale), with QTL regions identified for seedling traits: second leaf length (LSL), second leaf sheath length (LSLS), second leaf blade length (LSLB), second leaf width (WSL), and LSL/WSL ratio

Full size image

Candidate Genes for the Second Leaf Morphology

Within the localized QTL regions associated with the second leaf morphology of triticale seedlings, eight candidate genes with high confidence were in silico identified (Table 3). Four candidate genes encoding the predicted (1) F-box/FBD/LRR-repeat protein, (2) myosin-10-like protein, (3) thylakoid membrane protein TERC as well as (4) xyloglucan endotransglucosylase/hydrolase were localized in 5R genome regions associated with LSL, LSLS, and LSLB (1), with LSL, LSLS, LSLB, and LSL/WSL ratio (2, 3) as well as with LSL/WSL ratio (4). The other candidate gene-encoding signal-peptide peptidase like (5) was localized on 6R region associated with WSL. The next two genes encoding pentatricopeptide repeat-containing chloroplast protein (6) as well as receptor-like cytoplasmic kinase (7) were localized on 4R region, associated with WSL. The last gene-encoding fertilization-independent endosperm 1 protein (8) was identified on 4R region, associated with LSL/WSL ratio (Table 3). In addition, extended STRING analysis of gene interactions with identified candidate sequences (Table 4) indicated two more candidate genes encoding: (9) WSL-related sec61 subunit beta, necessary for protein translocation in the endoplasmic reticulum and (10) LSL, LSLS, LSLB, and LSL/WSL-related DUF1995 domain-containing protein (Table 4). Functional enrichments in the above networks revealed their shared involvement in protein export as well as protein processing in endoplasmic reticulum and phagosome (KEGG: map03060, map04141 and map04145, respectively) as presented in Figure S3.

Table 3 Characteristics of the candidate genes identified within the QTL associated with the second leaf width (WSL), the second leaf length (LSL), the second leaf sheath length (LSLS), the second leaf blade length (LSLB) as well as with the second leaf length to width ratio (LSL/WSL) in DH ‘Hewo’ × ‘Magnat’ line seedlings of the winter triticale mapping population
Full size table
Table 4 STRING network data for the annotated interactive proteins, obtained via https://string-db.org/ tools
Full size table

Correlation Between Second Leaf Characteristics and Candidate Proteins Abundance

Next, in an attempt to verify candidate genes, the correlation values were calculated between the second leaf morphological traits of selected DH lines and the normalized abundance of proteins detected in the LC–MS analysis of the second leaf of these plants (Table 5, Fig. S4). The medium (R ≥ 0.5) and high (R ≥ 0.7) Pearson’s correlations were detected for aspartic proteinase/PA domain-containing proteins, NBS-LRR/F-box proteins, myosin heavy chain, protein tyrosine kinase, xyloglucan endotransglucosylase/hydrolase, pentatricopeptide repeat-containing protein as well as Low PSII Accumulation 3 DUF1995 protein and protein transport protein Sec61 subunit beta (Table 5). The highest correlations with the content of the above proteins were found for WSL and LSL/WSL measurements. The highest correlation values were between WSL and PA domain containing protein (− 0.79), NBS-LRR disease resistance protein (− 0.70 and − 0.75), as well as pentatricopeptide repeat-containing protein (− 0.78). In contrast to these negative correlations, LSL/WSL ratio was positively correlated with the abundance of the above candidate proteins (0.51, 0.58, 0.44, and 0.26 R values, respectively). This ratio was also positively correlated with other aspartic proteinases (R = 0.53, 0.67 and 0.71) and NBS-LRR disease resistance proteins (R = 0.58 and 0.45). In the case of three other NBS-LRR proteins, WSL was positively, while LSL/WSL was negatively correlated (Table 5).

Table 5 Significant Pearson’s correlations (R ≥ 0.5) found between the identified candidate proteins and individual characteristics of winter triticale seedlings
Full size table

Three morphological parameters: LSL, LSLB, and LSL/WSL ratio were positively correlated with myosin heavy chain content (R = 0.63, 0.73 and 0.62, respectively), while WSL was not correlated. The xyloglucan endotransglucosylase/hydrolase content was positively correlated with WSL (R = 0.65 and 0.64), while negatively correlated with LSL/WSL ratio (R =  − 0.33 and − 0.53). Contrary, Low PSII Accumulation 3 (LPA 3) protein content was negatively correlated with WSL (R =  − 0.50) and positively correlated with LSL/WSL ratio (R = 0.31). Seven asparic proteinases showed only positive correlations with morphological parameters related to length growth (LSL, LSLB, LSLS). F-box domain-containing protein, myosin heavy chain, and protein transport protein Sec61 subunit beta showed only positive correlations with all second leaf morphological traits (Table 5).

Discussion

The vitality and the development at the seedling stage are crucial in cereals’ life cycle, especially for winter cereals. It has been reported that rapid seedling establishment and early growth are important traits for improving yield (Von Arnim and Deng 1996; Aparicio et al. 2002). The morphology of winter cereals seedling is a trait that can be correlated with tolerance to freezing and fungal infection. Leaf length and width are important in spreading fungal pathogens and as a main assimilating organ, largely determine the size of the future yield (Nalborczyk et al. 1995; Aparicio et al. 2002; Kobayashi et al. 2003). The development of cereal seedling is a process controlled by metabolite and hormonal signals that participate in the gene expression, growth, and metabolism. Additionally, sugar regulation of the gene expression at the seedling stage may initiate a specific mechanism, which can participate in the control of this process.

In our study, analysed traits were selected based on their potential role in seedlings ability to overwinter as well as potential use as an indicator of tillering and future yield early evaluation. According to our knowledge, the present research is the first describing location of triticale genomic regions associated with seedlings leaf morphology, especially associated with the second leaf.

Among all identified regions, the most interesting ones were associated with the second leaf morphology, which were all located on rye chromosomes 4R, 5R, and 6R (Fig. 1). In this research, loci QTL_SL_4R associated with the second leaf width shown the highest phenotypic variation (20.6%) with the additive effect derived from ‘Hewo’ cultivar. On chromosome 4R, Miedaner et al. (2012) located QTL associated with plant height and 1000-kernel weight in rye. In triticale, on chromosome 4R, Dyda et al. (2019) located QTL regions of leaf damage and Fv/Fm chlorophyll fluorescence parameter both in hardened and non-hardened plants submitted to Microdochium nivale infection. Furthermore, in the same mapping population, Dyda et al. (2022b) identified one QTL region associated with B. graminis resistance together with a candidate gene associated with a race non-specificity and incomplete resistance protein. In the present experiments, within loci QTL_SL_4R, a candidate gene-encoding pentatricopeptide repeat-containing protein was identified (Table 3). Other authors’ research, conducted in Arabidopsis thaliana, showed that protein At4g18520 may be required for proper chloroplast development, the regulation of the plastid gene expression probably through regulation of plastid-encoded polymerase (PEP)-dependent chloroplast transcription, for RNA editing of several chloroplastic transcripts (especially accD transcripts), for processing of the chloroplastic rpoA pre-mRNA as well as for the monocistronic rpoA transcript processing (Yin et al. 2012). Also, Wąsek et al. (2022) in QTL region Qfr.hm-7A.1 a candidate gene At4g18520 was identified and associated with plant survival after freezing with the photochemical activity of the photosystem II. Other candidate gene, encoding fertilization independent endosperm 1 protein was located on chromosome 4R (Table 3). In maize, Fie1 is neither expressed in the sperm, egg cell, nor central cell before fertilization. Activation of the Fie1 maternal allele occurs around two days after pollination (DAP) in the primary endosperm and peaks at 10–11 DAP coinciding with endosperm transition from mitotic division to endoreduplication (Hermon et al. 2007).

In our study, the most significant QTL region was detected on chromosome 5R. Other authors indicated that chromosome 5R carries many important loci. Many studies described this chromosome regions associated with rye preharvest sprouting resistance (Masojć et al. 2007, 2009; Myśków et al. 2010; Milczarski et al. 2017). Also, in our previous research performed on the same DH ‘Hewo’ × ‘Magnat’ mapping population, nine QTL regions linked to morphological traits at the adult plant stage were detected on 5R (Wajdzik et al. 2019). Three of them, Qstl.hm.5R.1, Qph.hm.5R.1, and Qtkw.hm.5R.1 were related to the straw length, plant height, and thousand kernel weight, respectively and were detected in similar position to the regions described in this paper. In addition, the same LOD peak marker rPt-390522 at 7–8 cM position was indicated in Wajdzik et al. (2019) for the thousand grain weight as for the second leaf morphology in the present study. It might be a significant correlation between genomic regions detected for the second leaf length at the seedling stage with the straw length and the plant height at the adult stage. On chromosome 5R, Miedaner et al. (2012) localized loci for plant height and 1000-kernel weight in rye, which was also confirmed by Miedaner et al. (2018) who identified QTL regions on chromosome 5R for plant height at two growth stages. Myśków et al. (2018) identified on chromosome 5R loci associated with leaf rolling as the response of water deficit in a similar position to loci identified in our study. Furthermore, Dyda et al. (2019) identified loci associated with Fv/Fm chlorophyll fluorescence parameter in cold-hardened triticale seedlings inoculated with M. nivale strains in a similar position to loci described in this paper. The same LOD peak marker, 3040546 was identified on 5R at 30 cM position for LSL and LSL/WSL ratio herein and powdery mildew tolerance at adult plant stage described by Dyda et al. (2022b). Moreover, the same flanking marker, 4341499 was identified on 5R at 37 cM position for LSL/WSL ratio herein and powdery mildew tolerance in Dyda et al. (2022b) research. Flanking marker rPt-507500 identified in our present study for all traits was also detected for the triticale grain yield in Liu et al. (2016) research.

In addition in this study, few candidate genes were located among QTL regions identified on 5R chromosome and associated with multiple proteins (Table 3). Candidate gene-encoding thylakoid membrane protein TERC in Arabidopsis may function in several steps of PSII assembly but is especially indispensable during the early formation of the PSII reaction centre (Schneider et al. 2014). TERC has been also shown to interact with a PSII assembly factor ALB3 (Järvi et al. 2015). Also, on chromosome 5R gene-encoding xyloglucan endotransglucosylase/hydrolase was identified (Table 3). Increased protein abundance of xyloglucan endotransglucosylase/hydrolase (XTH) due to Cd exposure was observed, it promotes cell wall loosening and cell expansion in Medicago sativa (Gutsch et al. 2018). The LOW PSII ACCUMULATION 3 (LPA3) protein is probably located in the chloroplast stroma or associated with thylakoid membranes. LPA3 works with LPA2 as accessory proteins in the incorporation of psbC (also known as CP43) into PSII by interacting with the cpSRP ALB3 translocase. They are also believed to be involved in the synthesis and assembly of psbC (Järvi et al. 2015).

On chromosome 6R, loci identified in present study was in the same position as loci associated with plant height described by Miedaner et al. (2012). On the same chromosome, loci for preharvest sprouting were also identified (Masojć et al. 2007, 2009; Milczarski et al. 2017). Dyda et al. (2019) on 6R chromosome identified loci for Fv/Fm after triticale seedlings infection with three different M. nivale strains. In addition, loci for leaf rolling in rye (Myśków et al. 2018) and yield components (Börner et al. 2000) were identified also on 6R chromosome. Furthermore, Dyda et al. (2022b) identified four loci linked to B. graminis resistance in ‘Hewo’ × ‘Magnat’ mapping population together with seven candidate genes found in those regions.

Two out of eight candidate genes identified in the present study had predicted location in chloroplast and role in chloroplast RNA modification (pentatricopeptide repeat-containing protein) as well as thylakoid membrane biogenesis and thylakoid formation in early chloroplast development (thylakoid membrane protein TERC). Since TERC protein is essential for de novo synthesis of PSII core proteins, its contribution may be one of the explanations for the correlation found in this study between the morphological parameters of the second leaf and the chlorophyll fluorescence parameters. Other candidate genes products potentially contribute to membrane cleaving (signal peptide peptidase-like protein), defence response to chitin (receptor-like cytoplasmic kinase), and cellular glucan metabolic process (xyloglucan endotransglucosylase/hydrolase). The next two genes may be involved in negative regulation of long-day photoperiodism, flowering (F-box/FBD/LRR-repeat protein) as well as negative regulation of flower development and vernalization response (fertilization independent endosperm 1 protein). The last candidate gene is connected in cytoskeletal motor activity (myosin-10-like protein), possibly involved in leaf shaping and development. Above protein abundance correlation revealed with the use of LC–MS analyses seem to confirm the identified candidate genes; however, further research is needed using their verification by qRT-PCR.

Up to date, this is the first paper describing QTL regions and candidate genes/proteins associated with the second leaf morphology in winter triticale seedling after and without cold-hardening process. Taking into consideration only, the most significant QTL regions described in this paper, loci strongly associated with selected morphological traits were identified only on rye chromosomes 4R, 5R, and 6R. The information of significant QTL regions associated with triticale seedling development together with candidate genes and coded proteins taking part in plant development process can assist molecular breeding programmes in marker-assisted selection (MAS) and genomic selection (GS).