Phenotyping 172 strawberry genotypes for water soaking reveals a close relationship with skin water permeance

Plant material

Three collections of strawberries were investigated. The first collection (‘cultivars’) amounted to 64 named genotypes comprising advanced breeding clones and commercial cultivars of the cultivated strawberry (Fragaria × ananassa Duch.). The second collection (‘F2’) comprised 76 individuals and was a segregating F2 of a cross between the parental line ‘USA 1’ (F. chiloensis) and the cultivar ‘Senga Sengana’ (F. × ananassa). The third collection (‘wild species’) comprised a total of 32 genotypes that belonged to 12 Fragaria species of the ‘Professor Staudt Collection’: F. × bifera, F. cascadensis F. chiloensis, F. iturupensis, F. mandshurica, F. moschata, F. nilgerrensis, F. nipponica, F. nubicola, F. vesca, F. virginiana, and F. viridis (Olbricht et al., 2021). All plants were grown in pots filled with a commercial growing medium (Substrate 5TerrAktiv; Klasmann-Deilmann, Geeste, Germany) in a greenhouse at Weixdorf, Hansabred, Germany (51°8′52″N, 13°47′50″E). The greenhouse vents were programmed to open when the temperature exceeded 18 °C. There was no active cooling. Under these conditions the temperature in the greenhouse was always maintained below 35 °C. All collections were sampled in 2022. Only the cultivar collection was also sampled in 2023. The individual genotypes for the three collections are listed in Table S1.

A set of transgenic strawberry plants, cv. Chandler, was also used. These plants contained antisense sequences of the polygalacturonase genes FaPG1 (line PGI-29), FaPG2 (line PGII-8) or both (line PGI/II-16). Transgenic ripe fruits displayed a more than 90% reduction on the mRNA level of the corresponding PG gene, and lower PG activity than the un-transformed Chandler wildtype (‘control’) (Paniagua et al., 2020). Additionally, fruits from the commercial cultivars Camarosa and Amiga, were evaluated. Plants were grown in 22 cm diameter pots containing a mixture of peat moss, sand and perlite (6:3:1, v:v:v), and cultured in a confined greenhouse equipped with a cooling system at the Institute of Subtropical and Mediterranean Horticulture (IHSM) in Malaga, Spain (36°40′23″N, 4°30′14″W). The maximum temperature did not exceed 30 °C under natural daylight. The growing conditions of the transgenic plants including the un-transformed wildtype in southern Spain hence differed from those of the three collections in Germany. Plants were cultivated following IPM guidelines for strawberry production (LaMondia et al., 2002; Handley & Dill, 2009). Care was taken when watering that the fruit surfaces remained dry throughout development. Mature fruit free from visual defects were sampled randomly from a minimum of two plants per genotype.

Water soaking characteristics

Water soaking was induced by incubating fruit individually in deionized water (one fruit per 100 mL). Fruit was forced underwater using a soft polypropylene foam plug. At regular time intervals–usually 2-h intervals up to 6 h–fruit were removed and blotted dry using soft tissue paper. The surface area affected by water soaking increases with time and is easily recognized by visual inspection due to its pale, watery and translucent appearance (Fig. S1). Water soaking was quantified using a five-point rating scale (Hurtado & Knoche, 2021). The rating scale was: score 0 = no WS; score 1 = <10% of the surface area water-soaked; score 2 = 10 to = <35%; score 3 = 35 to 60% and score 4 = >60% of the surface area water-soaked. Earlier studies established that the water-soaked area (a continuous variable) and the rating scores (a discontinuous score) were reasonably well-fitted by a linear model (Hurtado & Knoche, 2021). A Gompertz curve was fitted to the time course of WS for each replicate. The duration of the lag phase (‘time lag’) and the slope of the linear phase (‘U’) were estimated from the regression equation (Fig. 1A). In practical terms, the ‘time lag’ represents the time until the first WS symptoms appear and ‘U’ the relative increase in WS per unit of time.

Water uptake characteristics

Water uptake was quantified gravimetrically (Hurtado et al., 2021). Briefly, fruit were weighed, incubated in deionized water, removed from water, blotted, and reweighed before incubation continued. Weighing intervals were 2 h up to a total of 6 h, unless otherwise stated. Rates of osmotic uptake (F_f, kg s⁻¹) were calculated as the slope of a linear regression fitted through data for fruit mass (kg) vs. time (s). Surface area of the fruit (A, m²) was estimated from the fruit mass using the following equation (Hurtado & Knoche, 2023a) (Eq. (1)):

(1) $$A=5.0756\times {\left(mass\right)}^{0.6547}$$

Dividing F_f by the value for fruit A yielded the flux density (J, kg m⁻² s⁻¹). The permeance for water uptake (P_f, m s⁻¹), also referred to as the filtration permeability (House, 1974)–was obtained using Eq. (2):

(2) $$P_f={\displaystyle \frac{F_f}{A_{fruit}\cdot \mathrm{\Delta }\mathrm{\Psi }}\cdot {\displaystyle \frac{RT}{\rho \cdot \overline{V_w}}}}$$

In this equation, ΔΨ (MPa) equals the gradient in water potential between the incubation solution (0 MPa) and the water potential of the strawberry. Since the turgor of strawberries is essentially zero (Hurtado et al., 2021), the fruit water potential is effectively equal to the osmotic potential of the expressed juice ( ${\mathrm{\Psi }}_{\pi }$, MPa). The latter was estimated from the total soluble solids (TSS; Brix°) of extracted juice determined using a refractometer (PAL 1, Atago, Tokio, Japan) and Eq. (3) (Straube et al., 2024):

(3) $${\mathrm{\Psi }}_{\mathrm{\Pi }}=-0.3292-0.0400\times TSS-0.0088\times TS{S}^2+0.0002\times TS{S}^3$$

The parameters $\frac{RT}{\rho \cdot \overline{V_w}}$ in Eq. (2) are constants and represent the universal gas constant (R, m³ Pa mol⁻¹ K⁻¹), the absolute temperature (T, K), the density ( $\rho$, kg m⁻³) and molar volume of water ( $V_w$, m³ mol⁻¹). The permeance values for water uptake so obtained measure the permeability of the fruit skin to osmotic water uptake in each named genotype. Fruit skin permeance is useful for comparisons since it is independent of fruit size and the driving force (i.e., the water potential difference driving water uptake).

Skin color, position of achenes, and firmness

Skin color was quantified in the equatorial region using a spectrometer (CM-2600 d; Konica Minolta, Tokyo, Japan). The hue angle was calculated according to McGuire (1992).

The position of the achene relative to the fruit surface was found to be highly variable. Also, earlier studies established that the achene depression is a site of preferential microcracking of the cuticle and that microcracks impair the barrier property of the cuticle in water uptake (Hurtado et al., 2021). To assess potential relationships between the achene position and the susceptibility to WS, the position of the achene was assessed using the three point rating scheme according to the standardized phenotyping of strawberry in RosBREED (Mathey et al., 2013). In this scheme achenes sunken in the surface were rated score 1, those level with the surface score 2 and those protruding above the surface score 3. In addition, achene dimensions (length, width, depth) were measured using a digital microscope (option depth composition image (3D- image), VHX-7000; Keyence, Osaka, Japan). Achene length and width were quantified by image analysis (cellSens Dimension 1.18; Olympus Soft Imaging Solutions, Münster, Germany).

For phenotyping the achene position in the field, a method based on a silicone cast was developed. The cast was prepared by filling the achene depression on the fruit surface using a fast-curing silicone rubber (Silicone rubber, SE 9186 Clear; Dow Corning Corp., Tokio, Japan). The cast was allowed to set, then carefully removed from the skin and subsequently scanned using a digital microscope and measured as described above. A calibration curve between the cast and the respective achene depression (r² = 0.98***) was determined using 30 fruits with different achene depression characteristics, from protruding to very sunken depressions. The equation was (Eq. (4)):

(4) $$Dept{h}_{achene}=1.19\pm 0.03\times Dept{h}_{cast}$$

Fruit firmness was measured according to the standardized phenotyping of strawberry in RosBREED (Mathey et al., 2013). Fruit were gently pressed between forefinger and thumb and fruit firmness rated on an arbitrary three-point scale, where score 1 was firm, score 2 medium and score 3 soft.

Cuticle mass and strain relaxation

Cuticle characteristics were evaluated only in the ‘cultivars’ population since this collection had a sufficient number of fruit. Epidermal skin discs (ES) with one achene in the center were excised from the equatorial region of the fruit using a biopsy punch (4 mm diameter; Kai Europe, Solingen, Germany). The ESs comprised the cuticular membrane (CM), an epidermis, and adhering fragments of the subtending flesh. The CMs were enzymatically isolated following the procedure described earlier (Hurtado & Knoche, 2023b). Briefly, the ESs were incubated in a solution of pectinase (90 ml l⁻¹; Panzym Super E flüssig, Novozymes A/S, Krogshoejvej, Bagsvaerd, Denmark), and cellulase (5 ml l⁻¹; Cellubrix L; Novozymes A/S) buffered in 50 mM citric acid buffer. The pH was adjusted to pH 4.0 using NaOH. Sodium azide (NaN₃) was added at a final concentration of 30 mM to prevent microbial growth (Orgell, 1955). Four to six discs were excised per fruit. The isolation medium was refreshed once. Achenes were carefully removed. Adhering cellular debris was removed using a soft aquarelle brush and ultrasonication at 35 kHz for 10 min (RK 510; Sonorex Super, Bandelin electronic, Berlin, Germany). The CMs (n = 5 discs per rep) were dried above silica gel and weighed on a microbalance (M2P; Sartorius, Göttingen, Germany). Wax was extracted from the CMs by incubating the discs in CHCl₃:MeOH (1:1, v:v) for 24 h at room temperature. The dewaxed CMs (DCMs) so obtained were dried above silica gel and re-weighed. The amount of wax per unit area was obtained by subtracting the mass per unit area of the DCM from that of the CM.

The apparent biaxial strain release after CM isolation was quantified in a subpopulation of 10 genotypes, selected based on their contrasting susceptibilities to WS. The procedure described by Lai, Khanal & Knoche (2016) was used. Briefly, the hydrated isolated CMs were spread on a microscope slide and then flattened using a glass cover slip. Calibrated images were taken using a digital camera (Camera DP73; Olympus) and a binocular microscope (MZ10F; Leica Microsystems, Wetzlar, Germany). The area of the flattened disc after isolation (A_CM) was measured by image analysis (cellSens Dimension 1.18; Olympus Soft Imaging Solutions, Münster, Germany). The strain release was calculated from Eq. (5), where A_i represents the area of the excised disc and A_CM the area of the isolated relaxed CM.

(5) $$\mathrm{S}\mathrm{t}\mathrm{a}\mathrm{i}\mathrm{n}\mathrm{r}\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{a}\mathrm{s}\mathrm{e}\left(\mathrm{\%}\right)={\displaystyle \frac{A_i-A_{CM}}{{\mathrm{A}}_{CM}}\times 100}$$

A preliminary experiment was conducted to assess potential effects of the depth of the achene depression on the apparent biaxial strain release. Fruit were selected from genotypes having contrasting achene depressions ranging from very sunken to protruding. The ESs were excised as described above. The values of A_i were calculated as the surface of a truncated cone having an ellipsoidal base and a diameter equivalent to the diameter of the biopsy punch used for excision (diam. 4 mm). The value of A_CM was measured following incision and flattening of three hydrated CM discs on a microscope slide. Data analysis revealed there was no significant difference in strain release regardless of whether or not the shape of the achene depression was accounted for by the truncated cone model or whether the projected surface area (i.e., the cross-sectional area of the biopsy punch) was used to estimate A_i. Therefore, the latter was used to estimate the value of A_i.

Microcracking of the cuticle

Microcracking was assessed according to the procedure described by Hurtado & Knoche (2023b). Briefly, a fruit was incubated for 5 min in 0.1% (w/w) aqueous acridine orange (Carl Roth, Karlsruhe, Germany) and observed under fluorescent light using a binocular microscope (MZ10F; Leica Microsystems, Wetzlar, Germany). Calibrated images were taken in the equatorial region (of maximum diameter) (Camera DP73; GFP-plus filter, 460–500 nm excitation, ≥510 nm emission wavelength). Microcracking was indexed as the percentage of the surface area in the microscope field of view infiltrated by acridine orange. Since acridine orange does not penetrate an intact cuticle, the percentage of the surface area having orange, yellow and green fluorescence is a measure of the extent of cuticular microcracking (Peschel & Knoche, 2005).The area infiltrated with acridine orange was quantified using image analysis (cellSens Dimension 1.18; Olympus Soft Imaging Solutions, Münster, Germany).

Data analysis

All experiments were conducted using a minimum of 10 fruit sampled from a minimum of two plants per genotype. Data are presented as means ± SE. Where not shown, error bars were smaller than data symbols. Analysis of Pearson correlation, regression and variance was done using R (version 4.1.0; R Core Team, 2021).

Susceptibility to WS is closely related to skin permeance to water

Susceptibility to WS is closely related to the osmotic water uptake characteristics. Highly significant correlations were obtained in all three collections despite of their wide differences in genetic background and morphology. The consistency of these correlations suggests that the mechanism of WS is identical in all three collections investigated here.

It is unsurprising that correlations between susceptibility to WS and the water permeance of the skin were higher than those between susceptibility and water uptake flow rates (mass per time per fruit) or flux densities (mass per unit area) because permeance is independent of both fruit surface area (and hence of fruit size) and also of juice osmotic potential (and hence of the driving force for osmotic water uptake). Note that because strawberries lack significant turgor, the osmotic potential of the fruit’s juice is essentially equal to its water potential (Hurtado et al., 2021). The difference in water potential between the fruit and that of simulated rain in our WS induction assays equals the driving force in water uptake. Therefore, a relationship between the rate of water uptake and the juice’s osmotic potential was expected, but this was not significant.

The most likely explanation for both a fruit’s high susceptibility to WS and the high permeance of its skin is the presence of a large number of microcracks in the cuticle. Earlier studies established that strawberry cuticles are exceptionally thin and markedly strained (Hurtado & Knoche, 2023b). Strain results from the cessation of cuticle deposition at about color change when the genes involved in cuticle synthesis and deposition are downregulated. The increase in surface area that occurs thereafter essentially distributes a constant mass of cuticle over an increasing surface area. As a consequence of the growth stress, the cuticle is strained and microcracks form (Straube et al., 2024). That stress causes strain and strain causes failure is an established concept in material science (Niklas, 1992). Unfortunately, the strawberry cuticle is too thin to be isolated for mechanical testing. Thus, it is not possible to provide direct experimental evidence.

That the permeance of the skin is closely related to the susceptibility to water soaking is also consistent with the lack of significance or the low coefficients of correlation between WS susceptibility and either fruit mass or fruit osmotic potential.

Susceptibility to WS, microcracking and cuticle characteristics

Microcracks in the cuticle play a key role in determining susceptibility to WS. This conclusion is based on the following arguments. First, microcracks are sites of preferential water uptake, so fruit skins that suffer extensive microcracking are especially water-permeable (Hurtado et al., 2021). Second, the susceptibility to WS was significantly related to the skin permeance. Genotypes having a high skin permeance were most susceptible. Third, microcracking and WS in necked fruit occur especially in the neck; this zone has been identified as a zone of preferential water uptake (Hurtado & Knoche, 2023a).

Because cuticle properties such as mass per unit area, wax content and strain relaxation may affect the susceptibility to microcracking and microcracking in turn is related to WS, cuticle traits were phenotyped. Surprisingly, there was no relationship between a genotype’s susceptibility to WS and the strain release upon CM isolation, or the CM’s mass per unit area. The water permeance of the fruit skin was also not related either to CM strain or to CM mass per unit area. Unfortunately, the database for these difficult-to-measure properties was too limited for systematic phenotyping of the trait of CM fragility in a larger number of genotypes, such as in an entire genotype collection. The measurement difficulty is due to the extreme thinness of the CM, to the corrugated nature of the fruit surface and to the presence of achenes.

The roles of polygalacturonases and cell adhesion in WS

Polygalacturonases degrade homogalacturonan, which is the major constituent of the middle lamella that connects abutting cells (Jarvis, Briggs & Knox, 2003; Sénéchal et al., 2014). We expected the FaPG1 and FaPG2 transgenic lines to exhibit reduced polygalacturonase activity, and so suffer reduced homogalacturonan degradation, and so be less susceptible to WS than the ‘Chandler’ wildtype plants. However, this was not the case. There were no differences in WS susceptibility to the wildtype (‘Chandler’) when FaPG1 or FaPG1 and FaPG2 were down-regulated. The down-regulation of only FaPG2 resulted even in an increase in WS susceptibility. The latter observation could also be related to a significant decrease in cuticle thickness. Across mutants and cultivars, there was no consistent relationship between fruit firmness and susceptibility to WS. All transgenic lines had firmer fruit than the non-transformed ‘Chandler’ wildtype, but there was no reduction but even an increase in WS. The firmest fruit of ‘Amiga’, an (un-transformed) cultivar, was the least susceptible to WS. It is not known whether the sites of action of the polygalacturonases FaPG1 and FaPG2 are the same.

These findings should not be taken to imply that cell-to-cell adhesion is not a factor in WS. Water soaking is preceded by microcracking and by localized water uptake and this leads to the bursting of individual cells (Hurtado & Knoche, 2021). As a consequence, malic and citric acids leak into the cell wall space (Winkler et al., 2016). The permeability of the plasma membrane of adjacent cells increases causing further leakage. Malic and citric acids tend to extract Ca from the cell wall, which results in decreased cross-linking of homogalacturonans and other pectins. These processes will decrease cell-to-cell adhesion and thus facilitate the expansion of the water-soaked areas.