Identification and characterization of Gypsophila paniculata color morphs in Sleeping Bear Dunes National Lakeshore, MI, USA

Background Gypsophila paniculata (baby’s breath) is an invasive species found throughout much of the northwest United States and western Canada. Recently, plants exhibiting a different color morphology were identified within the coastal dunes along eastern Lake Michigan. The common baby’s breath (G. paniculata) typically produces stems that are purple in color (purple morph), while the atypical morph has stems that are green-yellow (green-yellow morph). The purpose of this study was to characterize these newly identified morphs and determine if they are genetically distinct species from the common baby’s breath in order to assess whether alternative management strategies should be employed to control these populations. Methods We sequenced two chloroplast regions, ribulose-bisphosphate carboxylase gene (rbcL), and maturase K (matK), and one nuclear region, internal transcribed spacer 2 (ITS2), from the purple morphs and green-yellow morphs collected from Sleeping Bear Dunes National Lakeshore, MI, USA (SBDNL). Sequences were aligned to reference sequences from other Gypsophila species obtained from the Barcode of Life Database and GenBank databases. We also collected seeds from wild purple morph and wild green-yellow morph plants in SBDNL. We grew the seeds in a common garden setting and characterized the proportion of green-yellow individuals produced from the two color morphs after 5-months of growth. Results Phylogenetic analyses based upon rbcL, matK, and ITS2 regions suggest that the two color morphs are not distinct species and they both belong to G. paniculata. Seeds collected from wild green-yellow morphs produced a significantly higher proportion of green-yellow individuals compared to the number produced by seeds collected from wild purple morphs. However, seeds collected from both color morphs produced more purple morphs than green-yellow morphs. Discussion Based upon these results, we propose that the two color morphs are variants of G. paniculata. Given the significant difference in the number of green-yellow morphs produced from the seeds of each morph type, we also suggest that this color difference has some genetic basis. We propose that current management continue to treat the two color morphs in a similar manner in terms of removal to prevent the further spread of this species.

relationship between the green-yellow morph and the common purple morph.  (ThermoFisher, Waltham, MA).

4
The DNA of both green-yellow morphs and purple morphs was amplified at three 9 5 proposed 'barcoding' genes for plants: large subunit of the ribulose-bisphosphate carboxylase 9 6 gene (RbcL), maturase K (MatK), and internal transcribed spacer 2 (ITS2). The RbcL region was base pair (bp) for RbcL, , and 1498 bp for the three regions bootstrap analyzes, as well as using neighbor joining, and parsimony models. We also For TCS, gaps were treated as missing data and we used a 95% connection limit. Our results indicate that the green-yellow morph identified in SBDNL is not a genetically haplotypes. These haplotypes group within the same network as the G. paniculata reference, suggests that both color morphs belong to G. paniculata.
Of the regions analyzed for the green-yellow morphs, purple morphs, and reference 1 5 0 sequences, RbcL was the most conserved sequence with an overall mean genetic distance (d) = purple G. paniculata morphs and one green-yellow morph that clustered together inside the G. The phylogeny constructed from the RbcL sequence showed one G. paniculata reference were obtained from the BOLD database, and it is possible that this specific individual was which is one characteristic of the annual G. elegans. While there is also sequence data for both 1 6 3 the ITS2 and MatK regions for this specimen in BOLD, these were not used for this dataset due 1 6 4 to higher levels of sequence fragmentation compared to our samples. However, both the ITS2 1 6 5 and MatK sequences from this G. paniculata reference sequence aligned most closely to G. While our data suggest that both the purple and green-yellow morphs are G. paniculata, whether the green-yellow morph is a different commercial genetic strain is still unclear. RbcL, MatK, and ITS2 are common 'barcode' genes used to delineate plant species, but their ability to collected. However, given the distinct grouping between the G. paniculata reference sequence 1 7 8 and our samples, ITS2 may be a promising region to assess the ability to distinguish among 1 7 9 different commercial strains or regional varieties of G. paniculata. If this is the case, then our 1 8 0 data would suggest that the purple and yellow-green morphs not only belong to the same species 1 8 1 (G. paniculata), but are likely derived from the same or similar initial source populations.

8 2
The mechanism driving the color difference between the purple and green-morphs is approximately an acre-sized area and interspersed throughout large groups of purple morphs.

8 7
Occasionally, small pockets of green-yellow individuals have also been found together in other dispersal patterns of the two morphs throughout the dunes, the color difference observed does 1 9 0 not appear to be solely environmentally driven, and likely has a genetic component. Potential Aal et al. 2006, Knievel et al. 2009). Further work will begin to elucidate the specific mechanism 1 9 4 9 influencing these color difference in invasive G. paniculata populations, as well as to explore 1 9 5 whether this color difference drives functional differences between the morphs.

9 6
In terms of management strategies, the purple and green-yellow morph are not distinct 1 9 7 species and are likely derived from the same genetic strain, so there does not appear to be a need 1 9 8 to manage these two morphs differently. One concern with the green-yellow morph initially noted by TNC removal crews was that the taproot tended to be more diffuse than the purple  Therefore, current management approaches for these populations should be maintained to control 2 0 5 the further spread of G. paniculata throughout the Michigan coastal dune system.