Origin

‘UC Golden Gate’ (17C139P045) is a clonally propagated individual from a full-sib family (17C139) developed by hybridizing ‘UCD Moxie’ (11C141P001) and 10C017P001 (Fig. 1). ‘UC Keystone’ (17C138P062) is a clonally propagated individual from a full-sib family (17C138) developed by hybridizing ‘UCD Moxie’ and ‘UCD Warrior’ (08C138P003; Knapp et al. 2019b; Fig. 1). These parents were selected because they are high yielding, are large and firm fruited, and have long shelf life; were discovered to be resistant to Fusarium wilt (California race 1) and heterozygous for FW1; and were predicted to be sources of favorable alleles for improving clone yield, earliness, and resistance to Verticillium wilt and Phytophthora crown rot (PhCR). ‘UCD Moxie’ is a day-neutral (PF/pf) cultivar with reduced runnering and high marketable fruit yields, e.g., the fruit yields of ‘UCD Moxie’ surpassed ‘Monterey’ by 52% and ‘San Andreas’ by 64% in coastal California yield trials (Knapp et al. 2019a). Apart from ‘UCD Moxie’ (released in 2018), ‘San Andreas’ was the only Fusarium wilt–resistant day-neutral cultivar developed for the California market that was available when the breeding and cultivar development work described here got underway in 2015 (Pincot et al. 2022).

View Figure

• Download as Powerpoint
• Download Figure

The parents of ‘UC Golden Gate’ and ‘UC Keystone’ were visually scored for runner and daughter plant growth (runnering) over three nursery growing seasons in Winters, CA, using an ordinal scale (1 to 5), where 1 = runnerless, 2 = reduced runnering, 3 = intermediate runnering, 4 = heavy runnering, and 5 = excessive runnering (Brukental et al. 2025). The runner score means ($\overline{y}$) reported here were estimated from linear mixed model (LMM) analyses using the ‘lme4’ and ‘emmeans’ R packages with entry as a fixed effect and year and entry $\times$ year as random effects (Bates et al. 2015; Lenth 2019). ‘UCD Moxie’ was classified as reduced runnering ($\overline{y} = 2.0$), whereas ‘UCD Warrior’ ($\overline{y} = 4.0$), 10C017P001 ($\overline{y} = 4.0$), and ‘Monterey’ ($\overline{y} = 4.0$) were classified as heavy-runnering (Brukental et al. 2025). ‘UCD Warrior’ and 10C017P001 were selected as parents for hybridization with ‘UC Moxie’ to develop cultivars with stronger runnering than ‘UCD Moxie’ but weaker runnering than ‘Monterey’.

The resistance of ‘San Andreas’ and susceptibility of ‘Monterey’ to Fusarium wilt, specifically to a California race 1 isolate of F. oxysporum f. sp. fragariae (AMP132), were unknown when those cultivars were released in 2008. Their resistance phenotypes and the resistance phenotypes of the parents of ‘UC Golden Gate’ and ‘UC Keystone’ were ascertained in parallel with the discovery of FW1 in genetic studies conducted between 2015 and 2018 (Table 1; Pincot et al. 2018, 2022). Using Kompetitive allele-specific primer (KASP) and 50K array markers for SNPs associated with the FW1 locus (Hardigan et al. 2020; Pincot et al. 2022), ‘UCD Moxie’, ‘UCD Warrior’, 10C017P001, and ‘San Andreas’ were predicted to be heterozygous (FW1/fw1), whereas ‘Monterey’, ‘UCD Royal Royce’, and ‘UCD Valiant’ were predicted to be homozygous for the recessive (susceptible) allele (fw1/fw1). Their DNA marker–predicted resistance phenotypes were confirmed by 3 years of screening clones artificially inoculated with AMP132 as described by Pincot et al. (2022). The KASP marker used for the initial analysis was FW1-K7 (A/C; bp 432,840). The 50K array SNP markers used for confirmatory analyses were AX-184326400 (T/C; bp 291,251), AX-184226354 (T/C; bp 414,365), and AX-184863196 (T/C; bp 1,012,042) (Hardigan et al. 2020). These SNPs flank the FW1 locus on chromosome 2B (Pincot et al. 2022). The physical positions of these SNPs and others reported herein were ascertained in the haplotype-phased ‘Royal Royal’ genome (FaRR1; https://phytozome-next.jgi.doe.gov/info/FxananassaRoyalRoyce_v1_0; Hardigan et al. 2020).

Table 1. Across-year estimated-marginal means (EMMs)ⁱ and t statistics for linear contrasts between EMMsⁱⁱ for disease symptom ratingsⁱⁱⁱ among test and check cultivars observed in 2019–20, 2020–21, and 2021–22 screening trials at Armstrong Farm, Davis, CA, USA.

View Fullscreen

‘UC Golden Gate’ and ‘UC Keystone’ were developed through a combination of phenotypic and marker-assisted selection between 2017 and 2022. They originated from two of 359 full-sib families (10,000 hybrid individuals) that were evaluated in the 2017–18 selection cycle, propagated from seed in the summer of 2017, and grown to maturity in Winters, CA, in 2017–18. Selected progenies were asexually propagated (cloned) in Winters, CA, in 2018–19 and Winters and Dorris, CA, each year thereafter until their official release in 2023. Clones (asexually propagated bare-root plants) for the first year of replicated testing (2019–20) were produced in Winters, CA. Clones for subsequent years of replicated testing (2020–21 and later) were produced in high-elevation nurseries (Dorris, CA) using standard open-field plant propagation practices and postharvest chilling treatments optimized for the day-neutral market segment. The clones were harvested in late October, trimmed, and directly planted in late October and early November of each growing season on commercial farms in Nipomo, Santa Maria, and Prunedale, CA, and on the Armstrong Plant Pathology Farm in Davis, CA.

The first round of selection was applied to seed-propagated individuals phenotyped in the spring and summer of 2018 to identify perpetual flowering (day-neutral) individuals with outstanding fruit size, firmness, symmetry, color, gloss, and visual appeal; to eliminate individuals with fruit defects and deformities; to predict marketable fruit yields; and to evaluate and eliminate individuals with inadequate or excessive runnering. The entire population was genotyped with the FW1-K7 KASP marker in the spring of 2018 to identify and select individuals predicted to be heterozygous or homozygous for the FW1 allele (FW1_). Collectively, 94 individuals with FW1/FW1- or FW1/fw1-predicted genotypes and runner scores in the $2.5<y \le 3.0$ range were selected for clonal multiplication and disease resistance screening in Davis, CA, and advanced testing on commercial farms in Nipomo, Santa Maria, and Prunedale, CA.

‘UC Keystone’ and ‘UC Golden Gate’, check cultivars, the 94 selected individuals, and 149 to 168 additional selected individuals were phenotyped for resistance to Fusarium wilt, Verticillium wilt, and PhCR over three disease-screening seasons (November to July) in Davis, CA. We collected and analyzed 31,404 resistance scores for these diseases to support statistical analyses, selection decisions, and disease resistance claims. These individuals were screened using bare-root plants (four clones/individual/year) that were artificially inoculated with a single isolate of a single pathogen, planted in chemically fumigated ground in November, and phenotyped for disease symptoms in late spring and early summer using previously described production practices, screening protocols, and pathogen isolates (Jiménez et al. 2023; Pincot et al. 2020, 2022). They were grown in soils that had been chemically fumigated to suppress soilborne pathogens and eliminate the confounding effects of diseases caused by nontarget soilborne pathogens within each pathogen-specific screen. Entries were arranged in randomized complete blocks experiment designs with four replications/entry (clones/entry). The R packages ‘lme4’ and ‘emmeans’ were used for LMM analyses of ordinal resistance scores with entry as a fixed effect and replicate, year, and entry $\times$ year as random effects (Bates et al. 2015; Lenth 2019). We used ‘emmeans’ to estimate entry estimated marginal means (EMMs). The EMMs of test cultivars (‘UC Golden Gate’ and ‘UC Keystone’) were compared with the EMMs of resistant and susceptible check cultivars using pairwise linear contrasts (EMM₁ – EMM₂), where EMM₁ is the mean for ‘UC Golden Gate’ or ‘UC Keystone’, and EMM₂ is the mean for one of the check cultivars. The null hypothesis (H₀:EMM₁ = EMM₂) was rejected when the t statistic for the linear contrast was greater than the critical value from a t distribution with a type I error probability <0.05.

The 94 individuals selected in the seed-propagated plant stage in 2017–18 were cloned for replicated testing over three growing seasons on commercial farms in Nipomo and Prunedale, CA. Selected individuals and check cultivars were grown in two 20-plant plots arranged in randomized complete blocks experiment designs. These yield trials were grown using the management practices, raised-bed configurations, plastic mulches, planting densities, planting dates, irrigation, fertilization and pesticide application practices and schedules, and harvest schedules of our cooperators. The number of harvests ranged from 40 to 61 across yield trials. The count and yield of marketable fruit per 20-plant plot were recorded at each harvest. To compare kg/ha yields across trials, the yield/plant was multiplied by the number of plants/ha within each location. We collected and analyzed 18,526 observations for these traits to support statistical analyses and selection decisions. Fruit from early and peak season harvests were sampled from every trial to phenotype individuals for firmness, total soluble solids (TSS%), and titratable acidity (TA%) using methods described by (Petrasch et al. 2022). Over three growing seasons, 2,273 phenotypic observations were collected for these traits to support statistical analyses and selection decisions. The R packages ‘lme4’ and ‘emmeans’ were used for LMM analyses with entry as a fixed effect and replicate, location, year, entry $\times$ location, entry $\times$ year, and entry $\times$ location $\times$ year as random effects (Bates et al. 2015; Lenth 2019). As before, pairwise linear contrasts were used to test for statistically significant differences between test and check cultivar EMMs for each trait.

To evaluate shelf life–associated fruit quality traits, early and peak season fruit were harvested from 100-plant plots on commercial farms in Santa Maria and Prunedale, CA, in 2021–22 (Table 2). Fruit samples were stored at 4 °C for 14 d postharvest (DPH) in 19-cm $\times$ 11.5-cm $\times$ 7-cm plastic clamshells designed to hold ∼454 g of freshly harvested fruit. Fruit samples were evaluated at 0, 7, and 14 d DPH for marketability; the incidence of molds caused by Botrytis cinerea or other fungal pathogens; and fruit weight, firmness, and TSS. The marketability of fruit was visually assessed using ordinal rating scale (1 to 10) comparable to the one proposed by do Nascimento Nunes (2015), where scores increase as marketability decreases (1 = fruit in perfect condition and 10 = unmarketable fruit). Fruit were stored undisturbed in one clamshell/cultivar for phenotyping fruit marketability and weight and in a second clamshell/cultivar for destructively phenotyping random samples of fruit for firmness and TSS at each time point. The statistical analyses of these data were identical to those described earlier with entry and DPH as fixed effects and location, entry $\times$ DPH, and entry $\times$ location as random effects. We used the ‘emmeans’ R package to estimate the linear and quadratic effects of DPH on each trait and pairwise linear contrasts for comparison between test and check cultivar EMMs (Lenth 2019).

Table 2. Across-year estimated-marginal means (EMMs)ⁱ and t statistics for linear contrasts between EMMsⁱⁱ for fruit quality traits among test and check cultivars grown on commercial farms in Nipomo and Prunedale, CA, USA, over three growing seasons (2019–20 to 2021–22).

View Fullscreen

Description

‘UC Golden Gate’ and ‘UC Keystone’ are day-neutral. They produced an uninterrupted supply of fruit from the start of harvest near the vernal equinox through the spring and summer months on farms in Nipomo (35.0°N) and Prunedale (36.8°N), CA (Figs. 2 and 3). Their harvest-by-harvest and cumulative marketable fruit yields show that they flowered continuously under the temperature and photoperiod ranges normally encountered in coastal California locations where day-neutral cultivars are typically commercially grown (Figs. 2 and 3). Although we ended harvest by Sep 1, ‘UC Golden Gate’ and ‘UC Keystone’ continuously produced fruit and were commercially harvested through late October or early November each year.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

‘UC Golden Gate’ and ‘UC Keystone’ are Fusarium wilt resistant. This was substantiated by 3 years of screening plants artificially inoculated with AMP132, a California race 1 isolate of F. oxysporum f. sp. fragariae (Table 1). They were selected by screening DNA samples of seed-propagated full-sib progeny for the presence of the dominant FW1 allele using the FW1-K7 KASP marker (Pincot et al. 2022). ‘UC Golden Gate’ and ‘UC Keystone’ were heterozygous for the KASP-SNP marker (A/C) and predicted to be heterozygous for FW1 alleles. That prediction was confirmed using 50K array-genotyped SNPs in LD with FW1 (Hardigan et al. 2020; Pincot et al. 2022). The 50K array SNPs used for this purpose bracket the FW1 locus: AX-184326400 (T/C; bp 291,251), AX-184226354 (T/C; bp 414,365), and AX-184863196 (T/C; bp 1,012,042) (Pincot et al. 2022). Using these DNA markers, ‘Monterey’, ‘UCD Royal Royce’, and ‘UCD Valiant’ were predicted to be homozygous for the recessive (susceptible) allele (fw1/fw1). Consistent with these DNA marker predictions, ‘UC Golden Gate’, ‘UC Keystone’, and ‘San Andreas’ were confirmed to be resistant, whereas ‘Monterey’, ‘UCD Royal Royce’, and ‘UCD Valiant’ were confirmed to be susceptible to Fusarium wilt (Table 1). The mean symptom ratings for ‘UC Golden Gate’ ($\overline{y} = 1.17$), ‘UC Keystone’ ($\overline{y} = 1.25$), and ‘San Andreas’ ($\overline{y} = 1.15$) were not significantly different from one another (Table 1). The replicates of these cultivars were mostly symptomless ($y = 1$). Conversely, we observed wilting, necrosis, and death among replicates of the susceptible check cultivars (Monterey, UCD Royal Royce, and UCD Valiant). The mean symptom ratings were significantly greater (more severe) for the susceptible checks than ‘UC Golden Gate’, ‘UC Keystone’, and the resistant check ‘San Andreas’ (Table 1).

‘UC Golden Gate’ and ‘UC Keystone’ are moderately resistant to Verticillium wilt and PhCR. This was substantiated by 3 years of field screening plants artificially inoculated with virulent isolates of the causal fungal pathogens (Table 1). Their resistance was found to be comparable to the most resistant modern cultivars identified to-date (Feldmann et al. 2024c; Jiménez et al. 2023; Mangandi et al. 2017; Pincot et al. 2020). Over 3 years of replicated testing, the mean symptom ratings were not significantly different among test and check cultivars for either disease (Table 1). The PhCR resistance of ‘UC Golden Gate’, ‘UC Keystone’, and the check cultivars was partly attributed to RPc2, a large-effect quantitative trait locus (QTL) on chromosome 7B (Jiménez et al. 2023; Mangandi et al. 2017). The favorable RPc2 allele is dominant and increases PhCR resistance and, using screening protocols and experimental conditions identical to those reported here, the RPc2 QTL was previously estimated to explain 28.7% to 39.7% of the additive genetic variance for PhCR resistance (Jiménez et al. 2023). Using a 50K array-genotyped SNP associated with the RPc2 locus (AX-184109190; A/G; bp 22,332,550), we discovered that ‘UC Golden Gate’ and ‘UC Keystone’ are homozygous for the SNP allele (A) commonly associated with the favorable RPc2 allele (Jiménez et al. 2023).

‘UC Golden Gate’ and ‘UC Keystone’ are among the highest yielding day-neutral cultivars documented to date (Feldmann et al. 2024b). The fruit yields for ‘UC Golden Gate’ and ‘UC Keystone’ were estimated from harvests over three growing seasons on farms where losses to diseases caused by soilborne pathogens were negligible to nonexistent (Figs. 2 and 3; Table 3). Fruit were harvested from the start of the season (close to the vernal equinox) through the end of August each year following once or twice weekly commercial harvest schedules (the number of harvests per trial ranged from 40 to 61). Across years and locations, ‘UC Keystone’ produced the highest cumulative marketable fruit yields (116,243 kg/ha), followed by ‘UC Golden Gate’ (107,299 kg/ha), ‘UCD Valiant’ (102,875 kg/ha), ‘Monterey’ (98,997 kg/ha), and ‘UCD Royal Royce’ (99,805 kg/ha) across locations and years (Table 3; Fig. 3). Although these across-trial yield differences were not statistically significant, the yields of ‘UC Keystone’ were significantly greater than ‘Monterey’ within five of six trials, whereas the yields of ‘UC Golden Gate’ were significantly greater than ‘Monterey’ within four of six trials (Table 3; Fig. 3). ‘UC Golden Gate’ generally produced slightly higher early season yields than ‘UC Keystone’; however, their cumulative marketable yields were not significantly different in five of the six trials (Fig. 2). These data established that the intrinsic fruit yields of ‘UC Golden Gate’ and ‘UC Keystone’ are comparable to and often greater than ‘Monterey’, ‘UCD Royal Royce’, and ‘UCD Valiant’ in the absence of Fusarium wilt. The yield advantages of ‘UC Golden Gate’ and ‘UC Keystone’ over Fusarium wilt–susceptible cultivars are obviously bound to be much greater on farms infested with F. oxysporum f. sp. fragariae.

Table 3. Within-environment estimated-marginal means (EMMs)ⁱ and t statistics for linear contrasts between EMMsⁱⁱ for cumulative marketable fruit yield (kg/ha) among test and check cultivars tested on commercial farms in Nipomo and Prunedale, CA, USA, over three growing seasons (2019–20 to 2021–22).

View Fullscreen

‘UC Golden Gate’ and ‘UC Keystone’ are long-shelf-life cultivars with the requisite fruit quality characteristics necessary for minimizing postharvest losses (Tables 2 and 4; Figs. 4 and 5). The fruit firmness, TSS, and TA percentage ranges were comparatively narrow among test and check cultivars; however, many of the differences were statistically significant (Table 2). TSS percentages were significantly greater for ‘UC Keystone’ (8.89) than ‘UC Golden Gate’ (7.84), whereas the TA percentages were greater for ‘UC Golden Gate’ (0.85) than ‘UC Keystone’ (0.77) and ‘Monterey (0.76) (Table 2). TSS was significantly lower for ‘UC Golden Gate’ and ‘UC Keystone’ than ‘Monterey’ (Table 2). The TSS/TA ratio (a perceived sweetness metric) was significantly greater for ‘UC Keystone’ (11.01) than ‘Monterey’ (10.44) and ‘UC Golden Gate’ (9.45).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

‘UC Golden Gate’ and ‘UC Keystone’ are firm-fruited and among the most firmly fruited cultivars documented to date in strawberry (Tables 2 and 4; Jiménez et al. 2024). This trait is one of the most important determinants of shelf life—perishability increases and shelf life decreases as fruit firmness decreases (Harker et al. 2000; Cordenunsi et al. 2003; do Nascimento Nunes 2015; Jiménez et al. 2024; Petrasch et al. 2022; Watkins et al. 1999). The fruit of ‘UC Golden Gate’ (303.7 g-force) were significantly firmer than ‘UC Keystone’ (260.6 g-force), and the fruit of ‘UC Keystone’ were significantly firmer than ‘Monterey’ (239.1 g-force). ‘UC Golden Gate’ fruit are especially resistant to physical damage caused by harvest and handling. The fruit firmness of ‘UC Golden Gate’ and ‘UC Keystone’ was partly attributed to the homozygosity of a favorable POLYGALACTURONASE1 allele (PG1-6A1) that increases fruit firmness (Jiménez et al. 2024). Their PG1-6A1 genotypes were predicted using 50K array-genotyped SNPs in LD with the PG1-6A1 locus on chromosome 6A. The SNPs used for this purpose were AX-184953741 (T/G; bp 27,253,733), AX-184023221 (A/C; bp 27,593,175), AX-184726882 (T/G; bp 27,671,092), and AX-184210676 (T/G; bp 27,676,285), and AX-184242253 (T/C; bp 27,888,596).

There are other important but more subtle phenotypic differences between ‘UC Golden Gate’ and ‘UC Keystone’. ‘UC Golden Gate’ has a more compact plant habit and slightly longer trusses than ‘UC Keystone’ but shorter trusses than ‘UC Monarch’ (Cole et al. 2025). ‘UC Keystone’ has heavier vegetative growth and a denser plant canopy than ‘UC Golden Gate’. Consequently, ‘UC Golden Gate’ appears to be better adapted to tabletop, substrate, and tunnel production than ‘UC Keystone’, whereas ‘UC Keystone’ appears to be better adapted to organic production than ‘UC Golden Gate’.

The long shelf life of ‘UC Golden Gate’ and ‘UC Keystone’ was substantiated by evaluating fruit for changes in marketability (visual appearance), mold incidence, and fruit firmness, TSS, and weight over 14 DPH of cold storage (Figs. 4 and 5; Table 4). We did not observe molds caused by B. cinerea or other fungal pathogens at 0 or 7 DPH among the fruit samples tested (Fig. 4). The mold incidence was <0.7% at 14 DPH. Statistically significant differences were observed among cultivars for TSS and firmness but not weight or marketability. Moreover, the cultivar $\times$ DPH interaction was not statistically significant for any trait. We did not observe statistically significant changes in fruit firmness and TSS as DPH increased (linear and quadratic regressions of fruit firmness and TSS on DPH were not statistically significant). Conversely, we observed a statistically significant linear decrease in fruit weight and quadratic decrease in marketability as DPH increased (Table 4; Figs. 4 and 5). Figure 5 shows that the marketability of ‘UC Golden Gate’, ‘UC Keystone’, and ‘Monterey’ decreased at an increasing rate as DPH increased (score increase as marketability decreases). The fruit of every cultivar became increasingly more desiccated, lighter, less glossy, and less marketable as storage time increased (Figs. 4 and 5). The marketability scores were perfect (1 on the ordinal scale) to near perfect for every cultivar at 0 DPH and increased to 4 or 5 by 14 DPH. Fruit with scores in the 4 to 5 range are marginally marketable (do Nascimento Nunes 2015). These data show that ‘UC Golden Gate’ and ‘UC Keystone’ have a shelf life approaching 2 weeks under generic cold storage conditions but become increasingly less visually appealing and less marketable the longer the fruit are stored (Fig. 4).

Table 4. Effect of days postharvest (DPH) cold storage on fruit quality traits among test and check cultivars grown on commercial farms in Santa Maria and Prunedale, CA, USA in 2021–22ⁱ

View Fullscreen

‘UC Golden Gate’ and ‘UC Keystone’ deliver strong yields of bare-root plants in nursery production. Using an ordinal scale for scoring runner and daughter plant growth, Brukental et al. 2025 showed that ‘UC Golden Gate’ ($\overline{y} = 3.0$) and ‘UC Keystone’ ($\overline{y} = 3.2$) runner less strongly than ‘Monterey’ ($\overline{y} = 4.0$) and more strongly than ‘UCD Moxie’ ($\overline{y} = 2.0$). The runner scores for ‘UC Golden Gate’ and ‘UC Keystone’ are typical of most day-neutral cultivars and were not significantly different from the population mean ($\overline{y} = 3.0$) estimated from phenotypes observed in a study of several hundred elite and exotic generic resources, including heirloom and modern cultivars (Brukental et al. 2025). From a bare-root plant propagation perspective, they are an improvement over ‘UCD Moxie’ (reduced runnering). From a fruit production runner trimming perspective, they are an improvement over ‘Monterey’ (heavy runnering).

Availability

The release of ‘UC Golden Gate’ and ‘UC Keystone’ was approved by the College of Agricultural & Environmental Sciences at the University of California, Davis, in 2023. US Plant Patents have been granted for ‘UC Keystone’ and ‘UC Golden Gate’ (Knapp and Cole 2024a, 2024b). Plant breeder’s rights are pending in territories outside the United States. Those interested in acquiring plants of ‘UC Golden Gate’ or ‘UC Keystone’ for commercial purposes should contact the Strawberry Licensing Program at Technology Transfer in the Office of Research at the University of California, Davis (https://research.ucdavis.edu/technology-transfer/plant-variety-licensing-program/strawberry-licensing-program/).