Plant materials

‘Apple’ and ‘Tommy Atkins’ mangoes (Mangifera indica L.) were grafted on local, unclassified rootstocks. Fruit were harvested from a commercial orchard located in Mwala, Kenya (1°19’S, 37°26’E). The orchard was managed conventionally using local integrated crop management programs. Permission to sample fruit was given by the owner of the orchard Mr. and Mrs. Musyoka.

Fruit growth

The time course of fruit growth was established first. Fruits were sampled every 1–3 weeks from 41 to 159 days after full bloom (DAFB). Fruits were weighed (Sartorius Pro 32/34F micro scales, Sartorious AG, Göttingen, Germany) and their length, and two orthogonal diameters were measured using calipers (CD-30PK; Mitutoyo, Kawasaki/Kanagawa, Japan). Fruit surface area was calculated assuming spherical shape. A sigmoid regression model was fitted through a plot of surface area vs. time. From this model, the rate of surface area growth (cm² d^-1) was calculated from the first derivative.

The relationships between fruit diameter, fruit surface area and fruit mass were established. Briefly, three fruit per cultivar were sampled every 2 to 3 weeks and weighed (Sartorius Pro 32/34F). Calibrated digital photographs were taken from two orthogonal aspects (Lumix DMC-G80; Panasonic Corporation, Osaka, Japan). The three orthogonal fruit diameters were quantified using image analysis (ImageJ 1.53P; National Health Institute, Bethesda, MD, USA). Fruit were then peeled, the peels flattened between two glass plates and digital photographs taken (Lumix DMC-G80; Panasonic Corporation). Peel area was quantified using image analysis (ImageJ 1.53P).

To obtain information on the spatial distribution of the skin area increase over the whole fruit surface, a set of fruit was tagged. A square pattern of four dots of a non-phytotoxic white silicon rubber (RTV 744; Dow Corning, Midland, MI, USA) was printed in five different regions on the fruit surface. The regions were: stem end, cheek, apex, back and nak (see Fig 1). The dot pattern was photographed at regular intervals (Lumix DMC-G80; Panasonic) and the area ‘enclosed’ by the dots quantified (ImageJ 1.53P). When necessary, dots were reapplied. We calculated the relative area growth rate (cm² d^-1) from a sigmoid curve fitted through a plot of ‘enclosed’ surface area vs. time for each region.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Sketch of mango fruit showing the different regions of the fruit surface sampled.

These regions include stem end, cheek, apex, back and nak.

https://doi.org/10.1371/journal.pone.0258521.g001

Anatomy

Tissue blocks were excised from the fruit skin and the outer flesh, and incubated in Karnovsky fixative. These were stored in a cold room until use [9]. For processing, blocks were removed from Karnovsky solution and thoroughly rinsed with deionized water. Thin cross-sections of skin tissue were made in tangential periclinal direction in the cheek region of the fruit. The sections were stained for 5–8 min using 0.1% w/v calcofluor white (fluorescence brightener 28; Sigma-Aldrich Chemie, Munich, Germany), mounted on a microscope slide and transferred to the stage of a fluorescence microscope (BX-60; Olympus Europa Holding GmbH, Hamburg, Germany). Sections were examined under ultraviolet (UV) light (filter U-MWU 330–385 nm excitation, ≥420 nm emission; Olympus Europa) and calibrated images taken at ×20 magnification using a digital camera (DP73, Olympus Europa). The number of cells and cell dimensions (periclinal and anticlinal diameters) of the epidermal and three hypodermal cell layers were measured. Preliminary investigations established that epidermal and hypodermal cells are isodiametric in the tangential plane of the fruit as indexed by the absence of significant differences in their periclinal cell dimensions. There was no clear layering of hypodermal cells in either cultivar. We therefore measured the dimensions of all cells in each image (image size 349 × 262 μm) at 30, 50 and 70 μm depths from the surface of the cuticle. The periclinal area of a cell was calculated as the square of the cell’s periclinal diameter, the cell’s anticlinal aspect ratio as cell anticlinal diameter divided by the cell periclinal diameter. The number of cells per unit fruit surface area was counted.

Lenticel density and area determination

The epidermal segments (ES) were excised from the stem end, cheek, apex, back and nak regions using a biopsy punch (8 mm diameter). The ES were photographed (Lumix DMC-G80; Panasonic) and the number of lenticels per disc counted (ImageJ 1.53P). Lenticel density was established by dividing the number of lenticels per ES by the area of the ES. The number of replicates was 30.

The areas of individual lenticels were quantified using enzymatically isolated cuticles. Lenticels were viewed under bright incident light using a binocular microscope (Leica MZ10F; Leica Microsystems, Wetzlar, Germany). Calibrated digital photographs were taken (Olympus DP 71; Olympus Europa). The areas of three to four lenticels per CM were measured using image analysis (cellSens version 1.7.1.; Olympus Europa). The number of replicates was 60 lenticels from 10 fruits.

Cuticle deposition

Cuticles were isolated enzymatically. Briefly, ES were excised from the cheek region using a biopsy punch (8 or 10 mm diameter; Kai Europe, Solingen, Germany) at regular time intervals. The ES were incubated in an isolation medium containing pectinase (90 ml l^-1; Panzym Super E flüssig, Novozymes A/S, Krogshoejvej, Bagsvaerd, Denmark), cellulase (5 ml l^-1; Cellubrix L; Novozymes A/S) [10]. The pH was adjusted to 4.0 using NaOH. NaN₃ was added at a final concentration of 30 mM to suppress microbial growth. The solution was refreshed regularly until the cuticular membrane (CM) separated from adhering cellular debris. The CM were cleaned using a camel-hair brush, then rinsed at least 5-times using deionized water.

For mass determination, CMs were dried at 40°C (Memmert 100–800; Memmert, Schwabach, Germany) and then weighed (CPA2P; Sartorius). Wax was extracted in a Soxhlet apparatus using chloroform: methanol (1:1 v/v CHCl₃/MeOH) for a minimum of 2 h. The dewaxed cuticular membranes (DCMs) were dried overnight at 40°C (Memmert 100–800) and weighed. The CM, DCM and wax masses per unit area were calculated. The number of replicates was 10–20.

Quantifying the areas of ES, CM and DCM

The ES were photographed (Lumix DMC-G80; Panasonic) and their areas quantified by image analysis (ImageJ 1.53P). Following enzymatic isolation, the fully-hydrated CMs were again photographed using a binocular microscope (Wild M10; Leica Microsysteme; camera DP71, Olympus Europa). Before dewaxing, a square pattern of holes was punched into the center of each CM disc and the hole pattern photographed. This was necessary because the DCMs curled up following extraction and it became impossible to flatten them. Two orthogonal diameters were measured on each ES or CM disc. For the DCM discs the areas enclosed by the hole patterns were calculated.

Calculating apparent and cumulative strains

The developmental time course of change in apparent (ε′, %) and cumulative (ε, mm²mm^-2) strain of the skin was calculated as the strain released upon excision, and upon isolation, and upon wax extraction [7]. The apparent strains released after excision and isolation of the CM () and that released after dewaxing () were calculated using Eq 1 and Eq 2 respectively: (1) and (2)

In these equations, Aⁱ represents the area of the skin disc before excision; is the area of the isolated CM disc and is the area of an entire DCM disc. The latter was calculated from the square pattern of holes in the DCM. The sum of both these strains represents the total apparent strain (). As pointed out earlier, the area of the extracted CM disc at the particular time of sampling (t = i) serves as the basis of the calculation of apparent strains [7]. Any irreversible strain(s) that occur during growth of the carpel to the fruitlet at the time of sampling and that remain associated with the extracted CM discs are not accounted for.

These pitfalls are avoided when calculating the absolute cumulative strain (ε, mm²mm^-2) of developing fruit [7]. Here the increase in fruit surface area with time is accounted for. In this analysis, the total cumulative strain (ε_total) is partitioned into several component strains. The total cumulative strain (ε_total) at time t = i was estimated as the change in surface area (ΔAⁱ) relative to the initial surface area (A₀) at time t = 0 (Eq 3). As discussed earlier [7], it is impractical to define the starting surface area and time ab initio (i.e., the differentiation of the carpel within a bud). We therefore defined the surface area at the time of initiation A₀ somewhat arbitrarily as about 1 cm² and calculated the total cumulative strain from Eq 3 [7].

(3)

In this equation, represents the fruit surface area at a particular sampling time (t = i).

The cumulative strains released after excision and CM isolation (ε_exc+iso) and after dewaxing (ε_extr) were calculated as the differences in areas of the ES disc before excision (Aⁱ), and of the isolated CM (A_CM), and of the extracted DCM (A_DCM) (Eqs 4 and 5). Thus, Aⁱ equals the cross-sectional area of the biopsy punch corrected for fruit curvature. The A_CM and A_DCM were determined at each sampling time and then calculated as the A_CM and A_DCM on a whole-fruit basis at a particular time (t = i). This calculation assumes the amount of strain release is uniform across the entire fruit surface.

(4)

(5)

The strain that remains after dewaxing (ε_resid) is fixed mostly by the cutin matrix and is calculated from Eq 6 as follows: (6)

As explained earlier [7], the total cumulative strain (ε_total) represents the sum of these individual component strains because all were calculated relative to the same base (A₀) (Eq 7): (7)

Of the three component strains, ε_extr and ε_resid represent the plastic strain (ε_plastic), whereas the ε_exc+iso is the elastic strain.

Uniaxial tensile tests

Strips (5 mm wide) were excised from the CM isolated between 55 and 159 DAFB using parallel-mounted razor blades. Strips were mounted in a cardboard frame made from masking tape to prevent unintentional strain during handling. Strips were hydrated overnight by incubation in deionized water. Thereafter, strips were mounted in a universal material testing machine (clamping distance l₀ = 10 mm) (Z 0.5; Zwick/Roell, Ulm, Germany) equipped with a 10 N force transducer (KAP-TC; Zwick/Roell). Frames were cut open and the test initiated at a strain rate of 0.25 mm min^-1.

Stiffness (S) was calculated as the maximum slope of the force (F, N) vs. strain (ε_tensile, %) diagram. Uniaxial strain was calculated as the ratio of the applied strain (Δl) divided by the length of the relaxed sample, i.e., the clamping distance l₀ (mm) and multiplied by 100. The maximum force (F_max) and maximum strain (ε_max, %) represent the force and strain recorded at failure. Following testing, the fracture site was inspected on each specimen to identify whether or not the fracture was associated with a lenticel.

Data analysis and presentation

Data are presented as means and standard errors (SE). Where not visible, the SEs are smaller than the data symbols. Data were analyzed using analysis of variance with R statistical software (R version 4.0.3; R Foundation for Statistical Computing, Vienna, Austria) and SAS (SAS Institute, Cary, NC, USA). Means were separated using Turkey’s studentized range test (α = 0.05). Regression analyses were conducted in R (R version 4.0.3) and Sigma Plot (version 12.5; Systat Software, San Jose, CA, USA). Significances of regression equations at the 5, 1 and 0.1% levels are indicated by *, ** and ***, respectively.

Stress/Strain relaxation behavior of the cuticle reflects fruit growth and cuticle deposition

The stress/strain relaxation analysis reveals several differences between ‘Apple’ and ‘Tommy Atkins’. First, total cumulative strain was greater in ‘Tommy Atkins’ due to its larger fruit and, hence, greater fruit surface area (Figs 2 and 8). Second, the relationships between ε_exc+iso and ε_total (Fig 9A) and between ε_extr and ε_total were biphasic in both cultivars (Fig 9B). During early development, there was little change in ε_exc+iso and ε_extr as ε_total increased. The increase in ε_total was primarily accounted for by an irreversible increase in ε_resid (Fig 9C). Mechanically, the increase in ε_resid resulted principally from a deposition of cutin. In Malus apple, the addition of cutin to the inner side (cell-wall side) of the cuticle on the expanding fruit surface resulted in the fixation of elastic strain and, thus, the development of a radial strain gradient within the cuticle [11]. The incorporation of wax into the cutin network also causes a small degree of ‘fixation’ of elastic strain (reversible) and, hence, also contributes to the plastic strain (irreversible) [12]. Our results indicate the relationships earlier identified in Malus apple, also occur in ‘Apple’ and ‘Tommy Atkins’ mangoes. Here, the ongoing deposition of CM fixes the reversible elastic strains by converting them into irreversible plastic strains in both ‘Apple’ and ‘Tommy Atkins’. Accordingly, when the rate of CM deposition decreased, approaching an asymptote at 59 DAFB (‘Apple’; Fig 6C) and 68 DAFB (‘Tommy Atkins’; Fig 6D), during the second phase, ε_exc+iso and ε_extr both increased at higher rates as ε_total also increased (Fig 9). While, from a practical point of view, ε_exc+iso is truly elastic and reversible, ε_extr is a plastic strain in the cuticle on an expanding surface. However, mechanically, this strain is also reversible, since the strained cutin polymer relaxes when the wax that blocks the relaxation in vivo is extracted in vitro. It is particularly interesting that the breakpoint in the above relationships occurs earlier in ‘Apple’, as indexed by the lower ε_total, than in ‘Tommy Atkins’ (Fig 9). This observation is also consistent with the earlier breakpoint in CM deposition in ‘Apple’ as compared to ‘Tommy Atkins’. Furthermore, for any given value of ε_total, ε_exc+iso and ε_extr were consistently higher in ‘Apple’ than in ‘Tommy Atkins’ (Fig 9). This is in line with the lower rate of cutin deposition in ‘Apple’ than in ‘Tommy Atkins’ (Fig 6A and 6B). These arguments demonstrate that growth and cuticle deposition are the primary determinants of cuticle stress and strain. Both processes account for the differences between ‘Apple’ and ‘Tommy Atkins’ identified in the stress/strain relaxation analysis.

The cuticle of ‘Apple’ is mechanically weaker than that of ‘Tommy Atkins’

Uniaxial tensile tests revealed that the CM of ‘Apple’ is mechanically weaker than that of ‘Tommy Atkins’ (Fig 10). Stiffness, F_max and ε_max in ‘Apple’ CMs are all significantly lower than in ‘Tommy Atkins’ CMs. This implies that ‘Apple’ CMs fracture at lower force and lower strain than ‘Tommy Atkins’. The lower cutin mass (Fig 6A and 6B), the larger lenticels (Fig 5), and the greater strain release on excision of the ES and on isolation of the CM (Figs 7 and 9A) and on extraction of the CMs of ‘Apple’ (Figs 7 and 9B) may explain their relative weakness, compared with the CMs of ‘Tommy Atkins’ [13]. In addition, the presence of lenticels may also affect the mechanical strength of the CM. Microcracks in the CM are almost always initiated at a lenticel and these are more frequent in ‘Apple’ than ‘Tommy Atkins’ (Fig 5, Table 1). Also, as fruit surface area increases, lenticel area also increases, while the number of lenticels per unit area decreases (the total number of lenticels per fruit being constant). The increase in lenticel area occurs at a markedly higher rate in ‘Apple’ than in ‘Tommy Atkins’ (Fig 5). At maturity, lenticel area was about three-times greater in ‘Apple’ than in ‘Tommy Atkins’. The contribution of lenticels to the mechanical strength of mango skin is unknown. However, in the skins of grapes, the lenticels represent points of stress concentration and, hence, of weakness [14]. Furthermore, in Malus apple the cuticle-periderm boundary is the weak link, often causing skin strips to facture in this region when subjected to uniaxial tensile tests [15]. The larger lenticels of ‘Apple’ mango also imply a larger cuticle-periderm boundary and hence, a weaker skin. We observed cracking across or around lenticels in 70% and 59% of the CM strips of ‘Apple’ and ‘Tommy Atkins’, respectively (data not presented). This is consistent with the above hypothesis. Mechanically, the increase in area per lenticel in ‘Apple’ mango may also be interpreted as a mechanically weak periderm of the lenticel, that fails continuously during fruit expansion, thereby triggering a continuous formation of periderm.

The size, orientations and number of epidermal and hypodermal cells were generally similar in ‘Apple’ and ‘Tommy Atkins’ mangoes (Fig 4). In Malus apples, the more russet-resistant cultivars are characterized by having smaller and more compactly arranged epidermal and hypodermal cells [16]. Furthermore, the russet-resistant apple cultivars had smaller cells that were less variable in size than those of the more russet-susceptible cultivars [17]. However, there were no comparable and consistent differences between ‘Apple’ and ‘Tommy Atkins’ mangoes.

Lower rates of CM deposition and larger lenticels therefore predispose the ‘Apple’ mango skin to fracture when subjected to strain induced by normal growth, thereby accounting for its higher susceptibility to russeting. The cuticular fractures impair the barrier properties of the skin and trigger the formation of a periderm. This sequence of events is consistent also with the initiation of russeting at lenticels and the exacerbating effects of surface moisture on russeting in ‘Apple’ mango [3]. ‘Apple’ mango is not unique in this respect. Similar relationships have been identified in Malus apple [18–20]. Unlike susceptible ‘Apple’ mango, russeting in Malus apple is not initiated in the lenticels.

The trigger(s) that initiates the dedifferentiation in the hypodermis, the differentiation of a phellogen and the cell division in the phellogen to produce the phellem is not known. Potential candidates are a change in the O₂ or CO₂ concentrations in the internal atmosphere of the fruit or a local decrease (more negative) in water potential in the region surrounding the microcracks due to increase transpiration [20].