The diploid (2n = 2x = 14) strawberry species Fragaria vesca has been embraced as a convenient model system for genomic research. Its favorable attributes include a ~240 Mb reference genome, short generation time, small plant size, close kinship to commercial octoploid strawberry Fragaria × ananassa, and membership in the economically important Rosaceae family [1, 2]. Molecular, genomic, and biotechnological research in strawberry has made substantial progress with respect to in vitro transformation systems [3–5], genetic linkage maps [6–8], transcript analysis [9–11], and genomic sequencing [12–14]. In particular, the F. vesca variety known as ‘Hawaii 4’ (PI551572 = CFRA 197) has come to prominence due to its ease of Agrobacterium-mediated transformation and regenerability [4, 15], the generation of an inbred derivative ‘Hawaii 4x4’ (Slovin J., unpublished), and choice of the ‘Hawaii 4x4’ line as the source of the first complete strawberry genomic sequence [14]. As such, genetic protocol refinement for ‘Hawaii 4’ is an important priority.

The original PI551572 germplasm accession was collected in Hawaii in 1983 by R. Bringhurst, and was maintained by the National Clonal Germplasm Repository (NCGR), Corvallis, Oregon, under the local identification number CFRA 197. Agrobacterium-mediated transformation and regeneration was first reported by Haymes and Davis [15]. Those authors described a single transformant, selected on kanamycin and carrying neomycin phosphotransferase (nptII) and β-glucuronidase (gus) marker genes, and the segregation of these marker genes in a first (R1) generation progeny population. PI551572 had been utilized in this initial effort in part because of two favorable phenotypic traits: it was a so-called ‘Alpine’ or semperflorens form, meaning that it had a perpetual (day-neutral) flowering habit as contrasted to the seasonal (short day) flowering habit of most wild strawberry genotypes; and it was a runnering form, in contrast to non-runnering ‘Alpine’ varieties such as ‘Yellow Wonder’ (PI 551827), ‘Baron Solemacher’ (PI 551507), ‘Reugen’ (PI 551834), and ‘Alexandria’ (PI 602923). F. vesca is typified by red fruit; however, ‘Hawaii 4’, and ‘Yellow Wonder’ have yellow fruit.

Subsequently, PI551572 was found to be the most favorably responsive among fourteen F. vesca germplasm accessions evaluated for transformability and regenerability by Oosumi et al. [4], who established an optimized system that employed hygromycin resistance as the selectable marker. The study of Oosumi et al. [4] was aimed at establishing an in vitro system that would facilitate high throughput development of T-DNA insertion lines [16] in diploid strawberry. Our aim [17–19] and approach differed in that we sought to introduce RNAi constructs, and that our vector system [20] employed kanamycin resistance. As a prelude to future research, we also sought to know whether regeneration was occurring via a somatic embryogenesis or an organogenic pathway, a distinction that had not been specified by Oosumi et al. [4].

Here we report significant new findings regarding the response of ‘Hawaii 4’ to in vitro transformation and regeneration, including explicit evidence that regeneration under the employed conditions occurs via a somatic embryogenic pathway, and the first report that a substantial proportion of the primary transformants of this diploid line are tetraploid. We also report several variant leaf morphologies that are transgene-specific, and describe a simple means of differentiating the altered leaf morphology resulting from tetraploidy from those of the wild type and transgene-related variant forms.

Our detailed observations of the early stages of transgenic plant regeneration under the employed conditions clearly reveal a pathway of somatic embryogenesis, encompassing heart, torpedo, and dicotyledon stages. This finding contrasts with that of Oosumi et al., [4], who achieved transformation and regeneration of F. vesca ‘Hawaii 4’, but who described a regeneration sequence beginning with organogenic shoot formation and followed by rooting upon transfer to hormone-free medium. In one of their figures, the latter authors indicated the presence of “embryonic callus” [4: legend of Figure three], but did not employ the term “somatic embryogenesis” or explicitly specify the nature of their regeneration pathway in their report.

Somatic embryogenesis has been reported in the octoploid, cultivated strawberry [22–26]. Importantly, Husaini and Abdin [24] found that the regeneration pathway from leaflet explants in cultivar ‘Chandler’ was steered toward either somatic embryogenesis or direct shoot formation depending upon the concentration of just one key medium component: thidiazuron (TDZ). This research group then reported the optimization of TDZ concentration for the promotion of somatic embryogenesis in ‘Chandler’ [25], stating that this was the preferred regeneration pathway for their applied research purposes. Other factors found to promote embryogenesis in strawberry include a period of culture in the dark [22, 27], and cold treatment [27].

Future efforts toward optimization of post-transformation plantlet regeneration in the widely used ‘Hawaii 4’ variety of the diploid model species F. vesca will likely benefit from a targeted approach that seeks to optimize either an organogenic or a somatic embryogenic pathway, depending upon which pathway best serves the project needs. Under the conditions employed in the present study, we documented regeneration via the latter pathway, thereby establishing a defined baseline for future methodological enhancement of the somatic embryogenic approach. Obvious directions for such efforts would be to evaluate the dose-responsiveness of ‘Hawaii 4’ explants to TDZ in comparison to BA, and the effects of culture in darkness or under varying light regimes.

The frequent (~17%) occurrence of tetraploids among post-transformation regenerants was an unexpected outcome. Although numerous reports of Fragaria regeneration have appeared, none has yet reported ploidy changes. In our study, the occurrence of tetraploid transformants was not specific to any particular construct sequence, as tetraploids occurred among the transformants that yielded regenerants regardless of construct type. The detection of tetraploidy was an unanticipated, ad hoc observation, for which reason we did not examine the potentially causal affects of experimental factors such as vector system or cultural conditions. Therefore, we cannot separate the various aspects of the transformation and regeneration procedures, considered alone or in concert, as possible causal factors in the induction of tetraploidy based upon available data. However, several intriguing questions are suggested, as discussed below.

First, is ‘Hawaii 4’ particularly susceptible to the induction of tetraploidy, or is such susceptibility a generalized phenomenon in F. vesca, in diploid Fragaria, or in Fragaria in general? The fact that elevated ploidies following in vitro manipulations have not been reported previously in Fragaria suggests the possibility that the phenomenon we documented in ‘Hawaii 4’ may have at least an element of genotype- and/or taxon-specificity. Alternately or additionally, the employed vector and delivery system may have been a factor in the tetraploid outcomes. We employed a specific Gateway® vector [20], whereas Oosumi et al., [4] employed the pCAMBIA-1304 binary vector, while our study and theirs both employed an Agrobacterium-based delivery system. Oosumi et al. [4] did not report the occurrence of tetraploidy among their transformants; however, a careful examination of the four transformants that were photographically documented by these authors suggests to us that the transformant depicted in their Figure eight-b [4] may be a tetraploid, as suggested by the distinctive leaf morphology that we have shown to be indicative of tetraploidy.

The tetraploid transformants of ‘Hawaii 4’ that we examined all shared a distinctive leaf morphology, which manifested as a quantifiably altered ratio (B/A) of central leaflet to overall leaf width. Increased cell size is a widely documented and general consequence of ploidy elevation, at least from the diploid to the tetraploid level. It is possible that increased cell size alone may account for the disproportionate broadening (versus lengthening) of all three leaflets of the strawberry trifoliate leaf, the net result of which is that the broadening of the central leaflet is proportionately greater than the overall broadening of the trifoliate leaf. As yet we have not defined the tetraploidy-associated change in leaf morphology at the cellular level. We have, however, thoroughly documented the fact of tetraploidy, and have provided a simple morphological metric that allows its detection and distinction from diploid plants among regenerant transformants of ‘Hawaii 4’.

The recognition that tetraploidy occurs frequently and that it has a distinctive phenotype when it does occur in ‘Hawaii 4’ transformants will enhance the ability of researchers to identify mutant, non-tetraploid forms in mutagen-treated and/or transformed ‘Hawaii 4’ plants. In the present study, we obtained transformants with variously altered leaf morphologies, and on the basis of our characterization of the effects of tetraploidy were able to partition ploidy-related from construct-associated alterations in leaf morphology, even when both occurred in the same plant. Thus, the petiolule elongation associated with introduction of construct E08 was present in both diploid and tetraploid regenerants (Figure 5B), and the tetraploid form was clearly distinguishable by its widened leaflets (Figure 5B - right) and altered B/A ratio. Contrastingly, the ruffled leaflet phenotype associated with construct F10 was considerably magnified in the tetraploid, as compared with the diploid, form (Figure 5C).

A useful catalogue of morphological features was provided by Slovin et al. [28], pertaining to F. vesca inbred line 5AF7, a yellow-fruited and runnerless ‘Alpine’ genotype closely related to ‘Hawaii 4’, which itself is a runnering variety. This comprehensive phenotypic description was intended to provide a baseline to which other forms including derivative mutant forms could be compared. Petiole length was among the features described; however, no mention was made of petiolule length. The petiolule elongation displayed by E08 transformants reveals and provides definition to an additional variable trait to be found in Fragaria, thus adding to the useful trait catalogue contributed by Slovin et al. [28].

Another interesting leaf phenotype occurred in association with construct D09 (Figure 5D), and this phenotype was represented only in diploid regenerants. Here, the leaves were of slightly reduced size, mostly due to narrowing of the leaflets, giving the plant an overall gangly look. Finally, several regenerants carrying construct E10 exhibited a variable increase in leaflet number, from the usual three to four, five, or even seven, wherein adjacent leaflets were sometimes partially fused (Figures 5E, 6). Pentafoliate leaves are a defining feature of diploid Fragaria species F. pentaphylla, a form indigenous to the Tibetan region. However, in F. pentaphylla, the additional leaflets are quite small or vestigial, and are attached much lower on the petiole [29], while in E10 the additional leaflets are attached at more-or-less the same point as the normal three leaflets (Figure 6). The instability of the E10 variant form and its occurrence in only one confirmed independent transformant invokes the possibility of an insertional or other mutagenic cause. A somaclonal variant of cultivated strawberry (Fragaria × ananassa) variety ‘Redcoat’ with a similar phenotype was described by Nehra et al. [30].

The results reported here define opportunities to optimize the existing ‘Hawaii 4’ protocol by focusing on treatments that specifically promote somatic embryogenesis. The reported morphological metrics and descriptions will guide future transgenic studies using the ‘Hawaii 4’ model system by alerting researchers to the potential occurrence of polyploid regenerants, and to differentiating the effects on leaf morphology due to polyploidy versus transgenic manipulations. The results reported here raise many intriguing questions about the transformation system and host genotype employed in the present study. Was tetraploidy induced by some aspect of the regeneration system, or did the polyploids arise from polyploid cells pre-existing in the explant material? To what extent was the host genotype a contributing factor? For constructs associated with altered phenotypes, why did only a small proportion of regenerated, GFP + plants display the associated aberrant phenotype? Yet aside from the generation of biological and methodological puzzles that invite further study, the present investigation also yielded intriguing mutant lines displaying a remarkable spectrum of leaf morphology alterations and resulting from manipulations of previously uncharacterized genes, thereby opening new avenues to the characterization of novel gene functions.