The Impact of Cell Division and Cell Enlargement on the Evolution of Fruit Size in Pyrus pyrifolia

Abstract

• Background and Aims Dramatic increases in fruit size have accompanied the domestication of Pyrus pyrifolia. To evaluate the contribution of cell division and cell enlargement in the evolution of fruit size, the following study was conducted.

• Methods Three wild Pyrus and 46 cultivated Pyrus pyrifolia cultivars were selected to examine cell number/size at time of pollination and at time of fruit harvest. The period of cell division was estimated by logarithmic curve of the increasing pattern of cell number, and its correlations with maturation period and final fruit size were analysed.

• Key Results Final fruit size is directly related to the number of cells produced in the period immediately following pollination. Late-maturing cultivars are larger than earlier-maturing cultivars and this is due to an extended period of cell division.

• Conclusions The evolution of fruit size in P. pyrifolia has mainly resulted from shifts in the ability of cells to divide rather than to enlarge.

INTRODUCTION

Regulation of fruit size is of major importance in higher plant development (Gillaspy et al., 1993), and an economical factor for many horticultural crops, including Japanese pear. Knowledge of the factors that affect fruit size can be particularly helpful for selecting superior genotypes, establishing breeding programmes and in orchard management (Westwood and Blaney, 1963). Therefore, there have been numerous studies, from many aspects, aimed at understanding mechanisms of fruit growth and the genetic and cultural influences (Westwood et al., 1967; Hayashi and Tanabe, 1991; Scorza et al., 1991; Gillaspy et al., 1993; Higashi et al., 1999; Cowan et al., 2001; Jackson, 2003; Harada et al., 2005; Zhang et al., 2005a–c). Although actual size depends on the interaction between genetics and environment, the potential size of fruit is genetically determined. This could be explained by the fact that the cultivars in a given species display a wide range of fruit size when grown under the same conditions.

Although fruit-bearing species, including the genus Pyrus, are taxonomically diverse they share a common feature. Fruit from domesticated species, such as tomato, apple and pear, have often been tremendously enlarged over that normally found in the progenitor wild species (Laney and Quamme, 1975; Hayashi and Tanabe, 1991; Grandillo et al., 1999; Tanksley, 2004; Harada et al., 2005). Rubstov (1944) suggested that there are >20 species of occidental pears and 12–15 species of oriental pears in the world. In Japan, the majority of commercial pear cultivars are derived principally from P. pyrifolia ‘Nakai’ (2n = 34). It has been proposed that intra-population genetic variation within Japanese pears is smaller than within Chinese sand pears, although both kinds of pears belong to the same species, P. pyrifolia (Kikuchi, 1948; Teng et al., 2002a, b). Intriguingly, dramatic increases in fruit size and variation in fruit morphology have accompanied the domestication of P. pyrifolia. While fruit size of the wild Japanese pear is very small, fruit of cultivated Japanese pear display a wide range of sizes (Kajiura and Sato, 1990). For example, the putative wild ancestor of the cultivated Japanese pear bears fruit weighing only a few grams, while a single fruit from a modern cultivar may weigh up to 2 kg, a nearly 100-fold increase in weight (Fig. 1). Moreover, there is a tendency for later-maturing cultivars to have larger fruits than earlier-maturing cultivars in Japanese pear, peach and apple (Hayashi and Tanabe, 1991; Abe et al., 1993), but no general explanation or model for the genotypic differences in fruit size has been advanced.

Fig. 1.

Fruit size extremes within the species of P. pyrifolia. On the left is a fruit from wild P. pyrifolia ‘Yamanashi’, on the right is a fruit from cultivated P. pyrifolia ‘Atago’.

Because domestication occurred in prehistoric times, no one knows the actual evolutionary pathway by which wild species gave rise to plants with larger and variably shaped fruit. So what are the underlying genetic, molecular and developmental changes that permitted wild progenitors to produce the large, highly variable edible fruit associated with modern agriculture? The most likely scenario is that early humans selected for mutations associated with larger fruit and, gradually, enough large-fruited mutations accumulated to give rise to our present-day cultivars, although recent investigations in apple do not support the concept of human involvement (Luby et al., 2001). In tomato, genetic analysis of crosses between cultivated species and their wild relatives indicated that the domestication process involved mutations at a number of different genetic loci (MacArthur and Butler, 1938; Banerjee and Kalloo, 1989). Although the quantitative nature of fruit size variation has severely inhibited the use of classic Mendelian techniques to identify and characterize the individual gene mutations associated with fruit domestication, significant progress has been made recently by using quantitative trait loci (QTL) analysis (Lippman and Tanksley, 2001; Tanksley, 2004).

Generally, tomato size is a function of the number of cells within the ovary prior to fertilization, the number of successful fertilizations, the number of cell divisions that occur within the developing fruit following fertilization, and the extent of cell enlargement (Bohner and Bangerth, 1988; Gillaspy et al., 1993). Similarly, fruit growth in P. pyrifolia is characterized by an initial period of rapid cell division, followed by a long period of cell expansion, primarily by vacuolation (Hayashi and Tanabe, 1991; Jackson, 2003). Many studies focused on the two stages and suggested that cell number and cell size are very important factors determining final fruit size (Hayashi, 1960; Jackson, 2003; Harada et al., 2005). Actually, ovary development before pollination is also one of the important aspects of fruit development, because the cells in the pericarp of the ovary at pollination are the basis of the following cell division which is crucial in the determination of final fruit size. In tomato, the cell number of fruit at anthesis has been regarded as a determining factor of final fruit size (Bohner and Bangerth, 1998); however, little attention had been paid to this stage in P. pyrifolia. Moreover, studies on variation of fruit size within a species might be more helpful for understanding fruit evolution than within a genus (Izhaki et al., 2002). Consequently, to answer the question of how Japanese pear was changed by domestication in fruit size, the contribution of cell division and cell enlargement for the evolution of fruit size in P. pyrifolia were evaluated by examining cell number/size at time of pollination and at time of fruit harvest.

MATERIALS AND METHODS

Plant materials

Forty-six cultivated Japanese pear (Pyrus pyrifolia ‘Nakai’) cultivars and three wild pears, ‘Yamanashi’ (P. pyrifolia), ‘Toyotominashi’ (P. mikawana, offspring of P. pyrifolia and P. betulaefolia) and ‘Manshumameinashi’ (P. betulaefolia), used in this study are listed in Table 1. The 19-year-old trees were selected from the pear germplasm collection at Tottori University, Tottori, Japan in 2005. These trees, grafted on Pyrus betulaefolia Bunge rootstocks, were spaced 2·5 × 4 m apart and cultured with a leader system. They annually received routine horticultural care and were hand-pollinated with ‘Chojuro’ pollen at anthesis. Fruit were hand-thinned to one per spur and the period of thinning was dependent on the cultivar (Hayashi and Tanabe, 1991).

Table 1.

Plant materials used in this study and their characteristic fruit size and maturation period

No.	Cultivar	Fruit size	Maturation period
Domesticated cultivar (P. pyrifolia)
1	Akitzuki	Large	Late
2	Atago	Large	Late
3	Hanheungli	Large	Late
4	Imamuraki	Large	Late
5	Kansaiyichi	Large	Late
6	Kuroki	Large	Late
7	Mishirazu	Large	Late
8	Niitaka	Large	Late
9	Okusankichi	Large	Late
10	Oushuu	Large	Late
11	Sekaiichi	Large	Late
12	Senryo	Large	Late
13	Shinsetsu	Large	Late
14	Taiheiyo	Large	Late
15	Akibaei	Medium	Middle
16	Amanokawa	Medium	Late
17	Chojuro	Medium	Middle
18	Gold Nijisseiki	Medium	Middle
19	Hattatsu	Medium	Middle
20	Hatsushimo	Medium	Late
21	Heishi	Medium	Middle
22	Hokkan	Medium	Middle
23	Housui	Medium	Middle
24	Imamuranatsu	Medium	Middle
25	Kousui	Medium	Middle
26	Kamenashi	Medium	Late
27	Kikusui	Medium	Middle
28	Nijisseiki	Medium	Middle
29	Nekogoroshi	Medium	Late
30	Ninomiyahakuri	Medium	Middle
31	Osa Nijisseiki	Medium	Middle
32	Seikiryu	Medium	Late
33	Shinju	Medium	Middle
34	Shinkou	Medium	Late
35	Shinsei	Medium	Late
36	Yahatanishiki	Medium	Middle
37	Zuishuu	Medium	Middle
38	Inagi	Medium	Late
39	Ruishannashi	Medium	Late
40	Awayuki	Small	Middle
41	Rokugatsu	Small	Early
42	Shinsui	Small	Early
43	Taihei	Small	Middle
44	Wasekozo	Small	Middle
45	Wasetaicho	Small	Early
46	Yakumo	Small	Early
Wild cultivar
47	Yamanashi (P. pyrifolia)	Small	Late
48	Toyotominashi (P. mikawana)	Small	Late
49	Manshumameinashi (P. betulaefolia)	Small	Late

Large, large fruit cultivars (>500 g f. wt); Medium, medium fruit cultivars (200–500 g f. wt); Small, small fruit cultivars (<200 g f. wt).

Fresh weight of fruit, and cell number and cell length in the mesocarp

Fruit of each cultivar were sampled at pollination and at harvest. Ten fruits (or flowers) per cultivar were weighed and immediately preserved in formalin–acetic–alcohol (80 % ethanol : acetic acid : formalin = 90 : 5 : 5) for histological analysis. The measurement of cell number and cell length in the mesocarp was conducted according to Zhang et al. (2005b) and modified. Firstly, the fruit was cut along the equatorial region (Fig. 2). Then, mesocarp width was calculated from the difference between the longest width of the transverse section of fruit and core. Subsequently, a transverse slice of mesocarp was taken along the equatorial region and stained by rubbing softly with a cloth soaked in blue ink. The stained surface was observed under a digital HF microscope system (VH-8000, Keyence, Tokyo, Japan) and an image from a CCD camera displayed on a monitor. Cell length, as an indicator of cell size, was measured from the length of seven contiguous cells from the core to the fruit surface: from these the average cell length was calculated. Ten observation zones per section were measured. The cell number in the mesocarp along the equatorial region was then calculated by dividing the mesocarp width by average cell length, and this was taken as an indicator of total cell number per fruit. Alternatively, the flower was also cut along the equatorial region with a blade. Since the flower is too small to examine with digital calipers, mesocarp width was also measured under a microscope and the number of cells in the mesocarp recorded. Studies in pear and apple have suggested that there is an extremely high correlation between cell layers and fruit fresh weight (Nii, 1998). Therefore, cell number along the equatorial region of mesocarp could be regarded as an indicator of fruit fresh weight or fruit size.

Fig. 2.

A diagram of the transverse section of Japanese pear fruit to show the experimental sampling site.

Determination of period of cell division

To analyse of the relationship between fruit fresh weight, maturation period and period of cell division of fruit in Japanese pears, three large-sized cultivars (‘Atago’, ‘Niitaka’ and ‘Shinsetsu’), three medium-sized cultivars (‘Housui’, ‘Kousui’ and ‘Shinkou’), two small-sized cultivars (‘Shinsui' and ‘Yakumo’) and wild P. pyrifolia (‘Yamanashi’) were selected (Table 2). Ten fruits per cultivar were collected weekly after anthesis and the cell number of the mesocarp along the equatorial region was measured according to the method described above. In a previous study, it had been suggested that the duration of cell division was cultivar-dependent in Japanese pear (Zhang et al., 2005b). To estimate the length of cell division, therefore, the increasing patterns of cell number in the mesocarp in all cultivars were fitted by logarithmic curves. The critical point when the slope of the fitted curve was below 0·5 cells d⁻¹ was calculated according to Zhang et al. (2005b), and the period from pollination to the critical point was regarded as the period of cell division for each cultivar. The fruit weight and maturation period of nine cultivars were recorded when they were commercially harvested.

Table 2.

Cultivars used for evaluation of the relationship between period of cell division and fruit fresh weight and maturation date

No.	Cultivar	Fruit size	Maturation period (d)	Period of cell division (d)	Fresh weight (g)
1	Atago	Large	210	56	1200
2	Niitaka	Large	185	42	765
3	Shinsetsu	Large	190	53	860
4	Shinkou	Medium	180	31	468
5	Kousui	Medium	128	28	282
6	Housui	Medium	135	33	340
7	Yakumo	Small	110	25	156
8	Shinsui	Small	120	27	186
9	Yamanashi	Small	160	21	35·7

Statistical analysis

The correlations between cell number, cell length, maturation period and period of cell division vs. fruit fresh weight were analysed by linear regression at P = 0·05. To investigate the differences in cell number in the mesocarp between large-, medium- and small-cultivated cultivars and wild P. pyrifolia, the average cell number for each group was selected for analysis by Duncan's multiple range test (P < 0·05). To estimate the period of cell division in fruit, logarithmic curves were fitted to the data of cell number of the mesocarp along the equatorial region by Sigmaplot software (Jandel Scientific, San Rafael, CA, USA).

RESULTS AND DISCUSSION

Cell number at and after pollination vs. final fruit size

In Japanese pear, fruit development can be divided into three phases. The first phase is involved in the development of the ovary and the decision to abort or to proceed with further cell division, which is generally referred to as fruit set. Prior to pollination, the development of ovary tissues including pericarp is usually characterized by cell division. In the following phase, fruit growth is due primarily to active cell division in the mesocarp. In the present study, cell number at harvest was estimated and considered to be the result of cell division at the second stage of fruit development. The result showed that there is no significant positive correlation (r = 0·1044) between cell number at pollination and final fruit fresh weight, but there is a positive and significant linear relationship (r = 0·7634) between cell number at harvest and final fruit fresh weight (Fig. 3). A further analysis of distribution of cell number among cultivar groups with different fruit sizes revealed that there were significant differences in cell number at harvest between them and no differences were observed at pollination (Fig. 4). It indicated that the cell number in fruit was cultivar-dependent and primary determined during the second phase of fruit development with active cell division in P. pyrifolia.

Fig. 3.

Correlation between cell number of the mesocarp and fruit fresh weight at pollination and harvest.

Fig. 4.

Comparison of mesocarp cell number along the equatorial region at pollination and harvest in P. pyrifolia. The cultivars examined were divided into cultivated and wild type. The cultivated cultivars were further divided into large-, medium- and small-size groups. Large, large fruit cultivars (>500 g f. wt); Medium, medium fruit cultivars (200–500 g f. wt); Small, small fruit cultivars (<200 g f. wt); Wild, wild Pyrus pyrifolia. Values followed by different letters are significantly different at P < 0·05.

Period of cell division after pollination; maturation period vs. final fruit size

In a previous study of Japanese pear, the duration of cell division was found to be cultivar-dependent with later-maturing cultivars having a longer period of cell division than earlier-maturing cultivars (Zhang et al., 2005b). In the current study, a positive and significant linear relationship (r = 0·7843) was found when the maturation period was regressed against the period of cell division (Fig. 5). In addition, the maturation period of cultivars and period of cell division against final fruit weight were significantly correlated (r = 0·8379, 0·9739, respectively). Abe et al. (1993) also suggested that, in Japanese pear, there is a significant relationship of inheritance between maturation period and final fruit weight. As shown in Table 1, it can be concluded that later-maturing cultivars usually have larger fruit than earlier-maturing cultivars in Japanese pear, and this could be explained by a longer period of cell division (Fig. 5) and a greater cell number in the former (Fig. 4).

Fig. 5.

Correlation between fruit fresh weight, maturation period, and period of cell division of fruit in P. pyrifolia.

Cell enlargement vs. final fruit size

The third phase of fruit development begins after cell division ceases. During this phase, fruit growth continues, mostly as a result of cell enlargement, until the fruit reaches its final size. Surprisingly, there was no positive and significant relationship (r =− 0·0170) between cell size at harvest and final fruit fresh weight in P. pyrifolia (Fig. 6). Therefore, it is likely that cell division is more important than cell enlargement in determining the final fruit size. Many previous studies of various fruit-bearing species, such as apple, peach, melon and avocado, also supported this suggestion (Westwood et al., 1967; Scorza et al., 1991; Higashi et al., 1999; Cowan et al., 2001; Zhang et al., 2005b).

Fig. 6.

Correlation between cell length of the mesocarp along the equatorial region and fruit fresh weight at harvest in P. pyrifolia.

Harada et al. (2005) proposed that a combination of greater cell division capacity and an enhanced degree of cell enlargement are involved in the increase in Malus fruit size. However, the present results showed that it is not the cell number in the pericarp at pollination but the cell number in the mesocarp at the cessation of the period of cell division after pollination that is crucial for determining final fruit size in P. pyrifolia. Moreover, it is not cell size but cell number in the mesocarp that is critical in the determination of fruit size. A longer period of cell division and the greater cell number in later-maturing cultivars of P. pyrifolia are the main factors contributing to the reason why later-maturing cultivars usually have larger fruit than earlier-maturing cultivars. On the other hand, large-fruited cultivars, such as ‘Atago’, have until now been widely grown as gift-pears in Japan because the Japanese prefer large pears. Also, fruit size is considered to be one of the important characteristics in new cultivar selection in Japanese pear breeding programmes (Kajiura and Sato, 1990). Therefore, it seems that the capacity for cell division has consciously or unconsciously been paid more attention than cell enlargement during evolution of P. pyrifolia from the wild type to the modern large-sized pear cultivars. In most instances, domestication resulted in both a dramatic increase in fruit size and enhanced variation in fruit shape. Therefore, the main reason for the evolution of fruit size in P. pyrifolia might mostly come from the changes in the ability of cells to divide rather than to enlarge.

Final fruit size is the consequence of complex metabolic events that occur during fruit development (Cowan et al., 2001). A previous study about the spur characteristics and the relationship between carbon partitioning and cell number of mesocarp among cultivars with different maturation dates also support the above hypothesis (Zhang et al., 2005b). It showed that photosynthate availability is crucial for fruit growth especially during the period of cell division, and the number of cells in the mesocarp is closely correlated with final fruit size. In other words, fruit size is a function of cell number rather than cell size, and factors affecting the activity of the cell division cycle assume importance. Fortunately, the studies of factors involved in cell division have been greatly advanced during the past few years. For example, Cowan et al. (2001) proposed that isoprenoid metabolism may be important in the control of cell proliferation and may affect final fruit size, because several end-products of the isoprenoid pathway, such as phytosterols, cytokinin and abscisic acid, are potentially involved in the control of cell division, fruit growth and fruit size. Recently, a QTL that increased grain productivity by increased grain number in rice, Gn1a, has been successfully cloned, which is a gene for cytokinin oxidase/dehydrogenase (OsCKX2), an enzyme that degrades the phytohormone cytokinin (Ashikari et al., 2005). In tomato, a number of QTL studies have suggested that less than ten loci account for the majority of the changes in size and shape associated with tomato domestication/agriculture (Grandillo et al., 1999). Among the key loci controlling fruit size in tomato, large-fruit alleles of fw2.2 are associated with a higher mitotic index (especially in cortical tissue) during the cell division stage just after anthesis (Cong et al., 2002). Cloning of fw2.2 has shown that this locus codes for a negative repressor of cell division, with activity confined largely to the cell division phase of fruit development (Frary et al., 2000; Cong et al., 2002). An interesting feature of fw2.2 is that the mutations associated with changes from small to large fruit are in the promoter, rather than in the coding portion, of the gene. Changes in gene regulation, rather than protein function, have long been hypothesized to play a major part in evolutionary change, especially when morphological differentiation is concerned.

However, fruit trees, including P. pyrifolia, always have a long period with juvenile, big chromosomes and few mutants, making it extremely difficult to identify and clone fruit weight QTL in these species as has been done in tomato. Intriguingly, the present study showed that the most important phase during fruit development for determination of final fruit size is the period of cell division. The fact that QTL cloning in tomato and rice is successful also implied that similar QTL might exist in other species. Although a similar gene fw2.2 has been successfully cloned in P. pyrifolia, there is no difference in copies of this gene between wild and cultivated cultivars (Hisatomi, 2003). On the other hand, it has been proposed that gibberellin (GA) is closely related to cell division and cell enlargement during fruit development in Japanese pear (Hayashi and Tanabe, 1991; Zhang et al., 2005c). Unpublished data showed that cultivated P. pyrifolia cultivars could produce much more active GAs than the wild P. pyrifolia cultivar, particularly in the period of cell division. Furthermore, the present finding that changes in fruit size from small-fruited wild pears to large-fruited cultivated pears are correlated with large shifts in cell number suggests that shifts in the ability of cells to divide can cause sizable changes in final fruit size. Hence, it is reasonable to expect that further studies focusing on the factors involved in cell division, such as the biosynthesis of GAs and cytokinin, would explore the key steps for manipulating fruit size in P. pyrifolia and other species. The current work is a first step into an examination of the variation in fruit size within species. Future studies will be needed to tease out the genetic and environmental controls on the variation in fruit size, and determine how tightly linked the traits are.