Effect of Cytoplasm Types T and D on Quantitative Trait Loci for Chip Color and Proline Content in Potato Tubers in a Diploid Potato Population

Abstract

The production of chips is an increasing part of the potato market. While the potato tubers are stored at low temperatures to minimize storage problems, they tend to accumulate reducing sugars, which negatively impact the quality and color of fried products. The goal of this study was to analyze the impact of cytoplasm type on chip color after harvest and after cold storage at 4 °C, as well as on proline content in cold-stressed potato tubers in a diploid potato population obtained from reciprocal crossing of parents with T- and D-type cytoplasm. Using 224 F1 progeny clones genotyped with Diversity Array Technology (DArTseq™), we mapped the Quantitative Trait Loci (QTL), treating cytoplasm type as a covariate. We detected five QTLs for chip color after harvest and six after cold storage, with the strongest QTL for both traits overlapping on chromosome III. Five QTL for proline content were detected on chromosomes V, X and XII, with the most significant one located on chromosome X. Although the progeny clones with T-type cytoplasm produced significantly lighter chips after cold storage, the cytoplasm type used as a covariate caused only minor modifications to the obtained QTL landscapes for chip color and proline content.

1. Introduction

Potato (Solanum tuberosum L.) is one of the most popular sources of staple food for the world’s population after rice (Oryza sativa L.), wheat (Triticum aestivum L.) and corn (Zea mays L.) [1]. There is a consistent trend of a decrease in fresh potato consumption and increasing consumption of processed potato, like chips and French fries, which is related to the changes in people’s lifestyles. The potato chips market constitutes a significant part of the potato market (even 37% in North America) and records annual increases [2].

The potato tubers are stored long-term at low temperatures, which limits losses caused by diseases, tuber shrinking (transpiration) or pre-mature sprouting. Unfortunately, the storage of tubers at low temperatures (4–6 °C) accelerates the conversion of tuber starch into reducing sugars, i.e., glucose and fructose. This phenomenon is known as cold-induced sweetening (CIS) [3]. Cold-stored potato tubers are unsuitable for thermal processing because, in the frying process, reducing sugars react with free amino acids in a Maillard reaction and generate dark-pigmented French fries and chips. Such products are bitter and visually unattractive to consumers [4]. In addition, one of the products of the Maillard reaction is acrylamide, which is a neurotoxin and a carcinogenic compound [5]. Hence, CIS causes considerable losses in potato processing [6], estimated to reach even 15% of stored tubers [7]. To reduce the negative impact of CIS, the potato processing industry reconditions the tubers for several weeks at 12–15 °C. Such a process is time-consuming and generates additional costs for the processors. For this reason, the best way to counteract CIS is to breed potato cultivars that are resistant to this phenomenon. CIS is genetically determined [8]. Quantitative trait loci (QTL) mapping revealed different chromosome regions that significantly affect the reducing sugar content/chip color in tubers after harvest and cold storage [9,10,11,12].

A factor that has not been taken into account in studies on CIS so far is the type of potato cytoplasm. The T-, D-, P-, A-, M- and W-cytoplasm types were found in the common potato gene pool [13]. Potato varieties available on the market and used in processing are mainly characterized by the T and D cytoplasm types. The D-type cytoplasm is derived from Solanum demissum, while the T-type cytoplasm was found in tetraploid Tuberosum potatoes [14,15]. Different types of potato cytoplasm have various impacts on the agronomic and quality traits, including the content of reducing sugars [16]. Jakuczun and Zimnoch-Guzowska [17] noticed a significant influence of crossing direction on glucose content in tubers obtained from reciprocal crosses. The authors showed a positive maternal effect on glucose content after storage at 4 °C and after reconditioning [17]. Knowing that the potato cytoplasm is inherited maternally, the demonstrated positive maternal effect of the cytoplasm on glucose content may indicate an influence of the cytoplasm on reducing sugar content and, therefore, CIS. What is more, there are known mutual, two-way signaling pathways between organelles and the nucleus [18]. In response to environmental changes, such as the stress of cold storage at 4 °C, signal molecules originating from organelles influence the expression of nuclear genes [19], thus directly influencing the plant phenotype. Taking the above into account, the cytoplasm may play a role in the accumulation of reducing sugars in potato tubers after cold storage.

In the response of tubers to cold stress, as in many other stresses, proline plays an important role, and the ability to accumulate proline under low temperatures may be crucial in a potato’s resistance to CIS. Proline is an osmolyte-type amino acid that has diverse functions in maintaining cellular homeostasis during growth and development in plants challenged by stress [20]. Proline regulates the osmotic balance in cells, stabilizes macromolecules and cell membranes, possesses free radical scavenging potential, and, thus, maintains redox balance. It is considered a biomarker for plant cold adaptation [20]. Its metabolism is under complex temporal and spatial control. Proline biosynthesis and degradation are catalyzed by enzymes located in the cytosol, chloroplast and mitochondria [21,22]. Many findings pointed to the role of proline in the response of plants to extreme temperatures [23,24]. In potatoes, proline content and proline metabolism-related gene expressions were important under drought stress [25,26]. To date, there are no reports on the genetics of proline accumulation in potato tubers.

The goal of this study was to analyze the impact of T and D cytoplasm type on chip color after harvest and cold storage and on proline content in cold-stressed potato tubers in a diploid potato population obtained from reciprocal crosses of parents with T- and D-type cytoplasm.

2. Materials and Methods

2.1. Plant Material and Assessment of Chip Color

Parental clones DG12-3/54 (T-type cytoplasm) and DG11-313 (D-type cytoplasm), and F1 progenies of two diploid potato reciprocal crosses (population T, N = 104; population D, N = 120), are described in [27]. Three plants (replications) per genotype were grown in semi-controlled conditions, in tents in the years 2019–2021, as described in [10]. Tubers of parents and progeny plants were evaluated for chip color after harvest (AH) and after 3 months of cold storage at 4 °C (CS). From each of the three plants (replications) of potato genotype, one tuber was selected for frying, and four slices, about 1 mm thick, were fried per tuber in oil at 180 °C for about 2–3 min. Visual assessment of color on a scale from 1 (dark) to 9 (light) was performed as described by Jakuczun et al. [28] for slices of each tuber separately.

2.2. Proline Content Assessment

Populations T (98 F1 individuals) and D (115 F1 individuals) collected in 2021 were evaluated for proline content (PC) in tubers after CS. One tuber from each of the three plants per genotype was selected and cut into slices by a standard kitchen slicer. The slices were combined, and 0.4 g of the material was ground in liquid nitrogen. Two independent PC measurements were performed from such prepared samples using a colorimetric method, according to Bates et al. [29].

2.3. Genetic Mapping and QTL Analysis

Total DNA was extracted from 200 mg of fresh, young leaves using the DNeasy Plant Mini Kit (Qiagen, Hilden, Germany). The amount of DNA and its quality were checked by NanoDrop Lite Spectrophotometer (Thermo Scientific, Waltham, MA, USA) and gel electrophoresis.

DNA samples of the parental and progeny clones were analyzed by a genotyping-by-sequencing approach (DArTseq™) at Diversity Arrays Technology Pty. Ltd. (Canberra, Australia; http://www.diversityarrays.com, accessed on 25 November 2024) as described by [30]. For genetic map construction and QTL mapping, the populations T and D were analyzed jointly as a mapping population of 224 individuals. DArTseq™ yielded 37,092 markers. The nucleotide sequences of the markers were located in the potato reference genome DM1-3 v6.1 [31] by a BLAST search. Resulting positions were encoded in the marker names by adding ‘c’ followed by chromosome number from 01 to 12 and position (Mb). Markers of presence and absence variant (PAV) type, scored as 0 or 1, were used for map construction. Markers that (i) did not segregate (showed value ‘1’ for less than 23 or more than 101 individuals), (ii) had missing data for parent(s) or value ‘0’ for both parents, (iii) had missing data for 8 or more progeny individuals, and iv) showed reproducibility < 1 were not included in the analyses. After import into JoinMap^® 4.1 [32], redundant markers with identical segregation patterns or similar to each other (similarity ≥ 0.95) were also removed.

The following JoinMap^® 4.1 settings were used for the construction of linkage maps by a regression method: CP (cross-pollination/outbreeder full-sib family) population type, independence LOD (significance cut-off, LOD score > 3) as a grouping parameter and the Haldane’s mapping function for the calculation of map distances. The DArTseq™ markers mapped to chromosomes other than in the reference potato genome (PGSC DM1-3 v6.1) were removed manually from the maps. Additionally, the 186 bp PCR marker ACA12 for the gene sequence of calcium-transporting ATPase 12, plasma membrane-type-like (Solanum tuberosum (potato)) (LOC107058116) was scored manually in the mapping population and added to the DArTseq™ dataset. This enzyme plays a role in sugar accumulation at low temperatures. The marker was amplified using forward primer 5′-CAATGGCTCCGCTACTGT-3′, and reverse primer 5′-CCGGTGGGATCTTGGTTGT-3′ as described previously [27]. In our previous study, in the population with T cytoplasm type, the expression level of calcium-transporting ATPase 12 was higher than in the population with D cytoplasm type [27].

QTL analysis was performed as described before [30], using interval mapping with MapQTL^®6 software [33]. QTL was detected using an LOD threshold ≥ 3.0 estimated from the cumulative distribution function of the maximum LOD on a chromosome for QTL analysis based on four QTL genotypes [34]. To test if AH, CS and PC are affected by the cytoplasm type, the cytoplasm types of the progeny were encoded as 0 (type T) and 1 (type D) and used as covariate/experimental design cofactor in the interval QTL mapping [33]. Then, we compared the results of QTL analyses obtained with and without the cytoplasm covariate.

3. Results

3.1. Chip Color and Proline Assessment

The two parental clones, DG12-3/54 and DG11-313, showed light chip color AH and CS, and their mean scores were 8.5 and 8.3 for AH and 8.2 and 7.7 for CS, respectively [27]. In the years 2019–2021, the mean values of chip color AH for the populations T and D were similar and reached 7.8 and 7.7, respectively. The range of variation of chip color AH for the T population was from 4.3 to 9.0 and from the D population from 5.1 to 9.0 (

Table 1. Distributions and fitness to normal curve of the phenotypic data: (a) mean chip color after harvest (AH) and after cold storage (CS) in years 2019–2021; (b) CS and proline content (PC) in tubers in 2021 (CS21, PC21) for populations T and D.

			T					D
Trait	N	Population Mean	Range of Variation	W Value Shapiro–Wilk Test	p Value	N	Population Mean	Range of Variation	W Value Shapiro-Wilk Test	p Value
(a)
AH	103	7.8	4.3–9.0	0.9385 *	0.0001	112	7.7	5.1–9.0	0.9632 *	0.0036
CS	103	7.4	5.5–8.8	0.9849	0.2931	119	7.1	5.0–8.5	0.9782	0.0504
(b)
CS21	99	7.8	5.1–9.0	0.9538 *	0.0016	113	7.4	4.5–9.0	0.9597 *	0.0018
PC21	98	15.9	5.1–52.8	0.8171 *	0.0000	115	17.6	5.0–69.5	0.7910 *	0.0000

* deviated from normality at p < 0.05.

a). The majority of F1 individuals within populations T (90%) and D (84%) had a light chip color (scores between 7.0 and 9.0) for AH (Figure 1). Similarly, as with AH, the majority of CS individuals in populations T and D showed light chip color, 68% and 53%, respectively (Figure 1). Chips fried after CS of some exemplary individuals from population T are shown in Figure 2. In the 2019–2021 dataset, population means for CS differed significantly between T (mean CS: 7.4) and D (mean CS: 7.1) populations (p = 0.002) according to the t-Student test for independent samples, where the grouping value is the type of cytoplasm. Values of CS were normally distributed in the mapping populations T and D, whereas AH values deviated significantly from normality, with a distribution skewed towards light-colored chips (Table 1a, Figure 1).

For 2021 data, the mean values of PC (PC21) for parental clones DG12-3/54 and DG11-313 were 26.8 and 28.5 µg/100 g of fresh weight (FW), respectively. For both populations, PC21 values for F1 individuals ranged from 5.0 to 69.5 µg/100 g FW. The most frequent PC21 scores (57%) ranged from 10 to 20 µg/100 g FW. Using datasets from 2021, we compared the two populations and while population means for PC21 did not differ significantly between T (mean PC: 15.9 µg/100 g FW) and D (mean PC: 17.6 µg/100 g FW) populations, a significant difference was detected between mean chip colors assessed in 2021 (CS21) of the T population (mean CS21: 7.8) and D population (mean CS21: 7.4; p = 0.011) according to t-Student test. PC21 in population T did not display a significant correlation (p = 0.208) with PC21 in population D at p < 0.05. The CS21 and PC21 distribution of the trait in populations T and D were evaluated for fitness to a normal curve by the Shapiro–Wilk test. For both traits, the distribution of the trait differed from normal (Table 1b). AH21 and CS21 did not display a significant correlation with PC21 in either of the populations at p < 0.05 (p value: 0.0328 for population T and 0.1046 for population D, respectively).

3.2. Genetic Mapping and QTL Analysis

The joint genetic map of both parents, DG12-3/54 and DG11-313, contained 2765 DArTseq markers and a PCR marker ATPase 12 (ACA12) (Supplementary Table S1). Among DArTseq markers, the nucleotide sequences of 2006 (72%) were found in the reference potato genome DM1-3 v6.1 [31], which allowed the identification and orientation of the 12 potato chromosomes. From 179 (chromosome X) to 324 (chromosome I), markers were mapped to the chromosomes, with an average number of 230 markers per chromosome (Supplementary Figure S1). The total length of the map was 996 cM, and the chromosome length varied between 58.1 cM (chromosome XI) and 131.8 cM (chromosome I; Supplementary Figure S1). The average chromosome length was 92.5 cM. The marker ACA12 mapped to chromosome IX at 62.0 cM, between the markers 5739016c0961 (located at 61,112,188 bp on chromosome IX bp in reference potato genome DM1-3 v6.1) and 3729499c0961 (61,482,094 bp). This corresponds well to the location of the nucleotide sequence on which it was designed (XM_015303453.1: 61,314,963–61,317,443 bp on chromosome IX of the PGSC DM1-3 v6.1).

3.3. QTL Analyses

The QTL for AH and CS detected by the interval QTL mapping in mean (2019–2021) datasets are shown in

Table 2. QTL for chip color after harvest (AH) and after cold storage (CS) detected by interval mapping in the mean dataset (2019–2021) of the joint mapping population of F1 diploid potatoes obtained from reciprocal crosses of parental clones DG12-3/54 and DG11-313 (a) and with the cytoplasm type used as a covariate (b).

Trait and Dataset	Chromosome	QTL Peak (cM)	Marker or Marker Interval at Peak	LOD	R2 (%)	Marker Origin a	Peak Position DM1-3 v6.1 b		QTL Range (cM)
							Chromosome	bp
(a)
AH	I	119.2	58168279c0184	3.48	7.1	P2	chr01	83,957,884	114.8–120.8
CS	II	38.5	3719288c0235	3.18	6.4	P2	chr02	35,196,853	38.3–38.5
AH	III	24.1	3688134c0339	6.71	13.4	P1	chr03	38,774,708	0.0–35.0
CS	III	12.9	5750470	6.60	12.9	H	NF c		0.0–65.1
AH	IV	7.3	3694144c041	3.04	6.3	P1	chr04	1,163,357	7.3
AH	IV	41.8	4450983c0456	3.01	6.2	P1	chr04	55,674,229	41.7–42.4
CS	IV	38.3	3723612c0443	4.93	9.7	P2	chr04	42,669,500	27.9–38.3
AH	X	52.6	5748594	3.17	6.6	P1	NF		52.6–53.2
CS	X	11.7	5714844	4.18	8.3	P2	NF		9.1–14.5
CS	X	52.3	5733100c1046	3.14	6.3	P1	chr10	46,488,902	51.6–52.3
CS	XII	12.1	3690083c123	3.69	7.4	P1	chr12	2,975,932	7.4–21.3
(b)
AH	I	119.2	58168279c0184	3.66	7.5	P2	chr01	83,957,884	114.6–120.8
AH	III	24.1	3688134c0339	6.70	13.4	P1	chr03	38,774,708	0.0–35.0
CS	III	12.8	3693458c037	6.47	12.0	P2	chr03	6,699,744	0.0–65.1
AH	IV	7.3	3694144c041	3.03	6.3	P1	chr04	1,163,357	7.3
AH	IV	41.8	4450983c0456	3.01	6.2	P1	chr04	55,674,229	41.7–42.4
CS	IV	38.3	3723612c0443	5.16	9.7	P2	chr04	42,669,500	21.1–45.1
AH	X	52.6	5748594	3.17	6.6	P1	NF		52.6–53.2
CS	X	11.7	5714844	4.11	7.8	P2	NF		9.1–14.5
CS	X	52.2	3724650	3.06	5.9	P1	NF		52.2–52.3
CS	XII	12.1	3690083c123	3.71	7.1	P1	chr12	2,975,932	7.4–21.3

^a P1—inherited from FR 313, P2—inherited from FR 54, H—descended from both parents of given population; ^b position in the reference genome S. tuberosum Group Phureja DM1-3 v6.1 [31] defined by BLAST search results; ^c not found in the reference genome.

. The QTL detected using datasets from single years of evaluation of AH and CS, are presented in Supplementary Table S2. Among the five QTL for AH, the strongest was the QTL on chromosome III, located at 0.0–35.0 cM, explaining 13.4% of the variance in the mean AH dataset (peak at 24.1 cM, LOD = 6.71;

Table 3. QTL for chip color after cold storage (CS21) and proline content (PC21) detected by interval mapping in datasets from the phenotypic assessments in 2021 of the joint mapping population of F1 diploid potatoes obtained from reciprocal crosses of parental clones DG12-3/54 and DG11-313 (a) and with the cytoplasm type used as a covariate (b).

Trait and Dataset	Chromosome	QTL Peak (cM)	Marker or Marker Interval at Peak	LOD	R2 (%)	Marker Origin a	Peak Position DM1-3 v6.1 b		QTL Range (cM)
							Chromosome	bp
(a)
CS21	II	38.5	3719288c0235	3.03	6.4	P2	chr02	35,196,853	38.4–38.5
CS21	III	34.1	4504190c0342	7.31	14.7	P2	chr03	41,576,761	3.1–65.1
PC21	V	46.6	3678154c0550	4.70	9.7	P2	chr05	50,090,721	40.9–54.6
CS21	X	51.6	3724239c1044	3.09	6.5	P1	chr10	44,437,803	51.6–52.1
PC21	X	9.1	3698078c101	4.13	8.5	H	chr10	618,308	7.5–9.1
PC21	X	36.7	58168117	6.26	12.7	P1	NF c		27.9–58.0
CS21	XII	18.3	3696170c123	3.66	7.6	P1	chr12	3,429,191	6.7–21.3
PC21	XII	46.8	3725047c1211	4.82	9.9	P2	chr12	11,433,467	46.1–47.2
PC21	XII	71.2	29296287c1256	3.99	8.3	P1	chr12	56,477,975	58.2–74.8
(b)
CS21	III	24.1	3688134c0339	6.83	13.4	P1	chr03	38,774,708	0.0–65.1
PC21	V	46.6	3678154c0550	4.96	10.1	P2	chr05	50,090,721	40.9–54.6
CS21	X	8.2	5739767c100	3.04	6.2	H	chr10	100,825	8.2–9.0
PC21	X	9.1	3698078c101	4.36	8.9	H	chr10	618,308	6.9–11.7
PC21	X	36.7	58168117	6.48	13.0	P1	NF		27.9–58.0
CS21	XII	6.7	3689672c122	3.63	7.4	P1	chr12	1,696,039	9.0–20.3
PC21	XII	46.8	3725047c1211	4.83	9.8	P2	chr12	11,433,467	46.1–47.2
PC21	XII	71.2	29296287c1256	3.90	8.0	P1	chr12	56,477,975	60.5–74.8

^a P1—inherited from FR 313, P2—inherited from FR 54, H—descended from both parents of given population; ^b position in the reference genome S. tuberosum Group Phureja DM1-3 v6.1 [31] defined by BLAST search results; ^c not found in the reference genome.

a). This QTL was significant in all three datasets from years of evaluation (AH19–AH21), with R² (%) varying between 9.8 and 11.2 (Supplementary Table S2). We detected six QTL for CS, of which the strongest was the QTL on chromosome III, spanning a wide region from 0.0 to 65.1 cM. This QTL explained up to 12.9% of the variance in CS (peak at 12.9 cM, LOD = 6.60; Table 2a), and it was significant in two datasets, CS19 and CS21 (Supplementary Table S2). The QTL for AH overlapped with QTL for CS on chromosome III (0.0–35.0 cM), and the QTL for both traits were located close to each other on chromosomes IV and X (Table 2).

To test if cytoplasm type affects AH and CS, the interval QTL mapping was repeated using the cytoplasm type as a covariate (Table 2b). The use of cytoplasm covariate affected only one of the five QTLs for AH. The QTL for AH on chromosome I became more significant, with an LOD increase of 0.18 and an R² increase of 0.4%. In the case of CS, a QTL on chromosome II (peak at 38.5 cM, LOD = 3.18, R² = 6.4%) became insignificant when the cytoplasm covariate was applied. The QTL for CS on chromosomes III and X decreased in significance (LOD decrease: 0.07–0.13, R² decrease: 0.4–0.9%), while the QTL on chromosomes IV and XII increased in significance (LOD increase: 0.02–0.23, R² increase: 0.0–0.3%), after application of the cytoplasm covariate. The peaks of QTL for CS on chromosomes III and X changed slightly their positions (Table 2b).

Five QTL for PC21 were detected on chromosomes V, X and XII. The most significant was the QTL for PC21 on chromosome X (27.9–58.0 cM), with the explained phenotypic variance of 12.7% (peak at 36.7 cM, LOD = 6.26; Table 3a). This QTL overlapped with QTL for CS21 (51.6–52.1 cM). However, this QTL for CS21 was not significant when the cytoplasm covariate was used in the QTL analysis (Table 3b). The cytoplasm covariate enhanced the significance of QTL for PC21 on chromosomes V and X, causing 0.22–0.26 LOD increase and 0.3–0.4% R² increase, while the two QTL on chromosome XII became weaker, with 0.1–0.3% R² decrease (Table 3). QTL for CS21 on chromosome III and QTL for PC21 on chromosome X, for which the most significant differences in LOD values analyzed with and without the cytoplasm covariate were detected, are presented in Figure 3.

4. Discussion

In potato tubers, the accumulation of reducing sugars is an outcome of complex pathways of starch biosynthesis and degradation influenced by many internal and environmental factors. To understand the CIS better, we analyzed two such internal factors: cytoplasm type and tuber proline content.

Cytoplasmic genome types influence various agronomic and quality traits in potatoes. Resistance to late blight and tuber bruising, male sterility, tuber shape, starch content and yield have been shown to be under cytoplasmic control [13,14,16]. In a previous study [27], using the same plant material, we compared transcriptomes of potato tubers and proteomes from amyloplast and mitochondrial fractions originating from potato tubers with T– and D-type cytoplasm and characterized by light and dark chip color after CS. Between pools of individuals differing in chip color after CS, there were 48 and 15 Differentially Expressed Genes (DEGs) for the T and D progenies, respectively. In the T-type cytoplasm, two amyloplast-associated and five mitochondria-associated Differentially Expressed Proteins (DEPs) were detected. Of 37 mitochondria-associated DEPs in the D-type cytoplasm, there were 36 downregulated DEPs in the dark chip color bulks. Analysis of the expression of the selected genes and quantity and quality analysis of the organelle-associated proteins have suggested that cytoplasm type might influence sugar accumulation in potato tubers [27]. In this study, we used a diploid potato population obtained from a reciprocal cross between parents with T and D cytoplasm types for QTL mapping. We used cytoplasm type as a covariate/experimental design cofactor in the interval QTL mapping to compare QTL data for AH and CS chip color obtained with and without the cytoplasm covariate. Although the progeny clones with T-type cytoplasm produced significantly lighter chips after cold storage than the progeny with D-type cytoplasm, the application of cytoplasm as a covariate resulted in only minor modifications of the obtained QTL landscape. This indicated a rather weak effect of the cytoplasm type on potato chip color and a low dependence of QTL for chip color on this factor.

Various studies mapped QTL for AH and CS on all 12 chromosomes of potato [9,10,11,12,15]. We detected five QTL for AH distributed over chromosomes I, III, IV and X. While the precise comparisons of the QTL locations between different studies are hampered by changing versions of the reference genome used over time, the QTL for AH detected on chromosomes I, III and X in our study seem to be in new locations, not described before [9,10,11]. The QTL for AH on chromosome IV overlaps with a QTL described by [11]. Of the six QTLs for CS detected in our study, the cytoplasm type-dependent QTL on chromosome II is novel. The remaining QTL appears to be approximately in the same genomic locations as reported previously on chromosome III [9], on chromosome IV [11] and on chromosome X [9,11]. The use of the same reference genome version DM1-3 v6.1 enabled a more precise comparison of the QTL locations between our results and the study of [12]. The QTL for CS located by [12] on chromosome III at 46.06 Mb in the reference genome overlapped with the position of our most important QTL for CS, spanning the region corresponding to 0–57 Mb in the reference genome. Similarly, on chromosome XII, the peaks of QTL for CS were close to each other at ca. 2.97 Mb (our study, Table 2) and at 7.44 Mb [12].

To withstand the cold, plants accumulate various compounds, including proline and soluble sugars [35]. In potato leaves, the level of proline increased with decreasing temperature [36]. In our study, we performed for the first time QTL mapping for proline content in potato tubers. Five QTL for PC21 were detected on chromosomes V, X and XII (Table 3a). There was no significant correlation between proline content and chip color, suggesting that both traits are determined by independent factors in potato tubers stored at low temperatures. However, QTL for PC21 and CS21 shared a similar region on chromosome X. Overlapping of QTL for different traits is expected if the observed phenotypes result either from pleiotropic effects of a single gene or from the effects of closely linked but otherwise unrelated genes [37]. In the absence of a covariate, QTL on chromosome X for PC21 ranging from 27.9 to 58.0 cM partially overlapped with the QTL for CS21 at 51.6–52.1 cM (Table 3a). In addition, in the presence of a covariate, we detected on chromosome X overlapping QTL for PC21 at 6.9–11.7 cM and for CS21 at 8.2–9.0 cM (Table 3b). The results may indicate the presence of common and cytoplasm-dependent regulatory factors for proline accumulation and chip color formation but with rather weak effects on the final chip color after cold stress.

Calcium (Ca²⁺) signaling plays a crucial role in plant cold responses, and Ca²⁺-ATPases (ACAs) localized in the plant cell membranes are involved in regulating cytoplasmic Ca²⁺ concentrations [38]. In our recent study, different expression levels of ACA genes were found in potato tuber samples characterized by light and dark chip colors after CS [27]. Here, the DNA marker ACA12 was mapped on chromosome IX, outside the QTL positions for AH, CS and PC. The result may indicate that this gene does not play a critical role in the regulation of reducing sugars and proline metabolism in potato tubers.

5. Conclusions

Previous data indicate the influence of potato cytoplasm type on the chip color after storing tubers at low temperatures. Our data show that chips obtained from T-cytoplasmic potato tubers are lighter than chips obtained from tubers with D cytoplasm type. Our research was performed on populations obtained from reciprocal crosses, which theoretically minimized the effect of nuclear genome diversity. QTL for chip color and proline content in potato tubers overlapped only partially. Against our expectations, QTL analysis with cytoplasm type as a covariate showed little effect of cytoplasm type on the detection of QTL for chip color and tuber proline content. Our results support the polygenic inheritance of chip color after cold storage and indicate cytoplasm type and proline content in tubers as minor modifiers of the trait.