Pitseed goosefoot seed collection expeditions in 2018

Personnel from Brigham Young University (BYU) made seed collection trips in 2018 to southern Texas (April), the New England coast (September), and Oklahoma (October). Seeds were sampled from populations encountered on publicly accessible roadside rights-of-way and public beaches, generally guided by previous collection sites noted on herbarium specimens. Maps of the three collections are included in Fig. 1. Collection site passport data and putative identification of C. berlandieri and accessions of other selected species were noted and are presented in Table 1.

Figure 1 Maps of BYU Chenopodium collection trips in 2018. (a) Southern Texas; (b) New England; (c) Oklahoma. 

Fruits from Chenopodium population samples were photographed at BYU using a BK PLUS Lab System (Dun Inc., Palmyra, Virginia, USA) with a 65-mm Canon (Melville, New York, USA) Macro lens at f4-5.6 with 1.5x-4x magnification. Images were captured using Capture One Pro 64 bit v. 10.2.1 (Phase One, New York, USA) and superimposed using Stacker v. 1.04 (Zerene Systems, Richland, Washington, USA), with contrast adjustment and sharpening performed using the Smart Sharpen tool in Photoshop x64 v. 12.0 (Adobe Inc., San Jose, California, USA).

DNA sequence analysis of C. pallidicaule, C. watsonii, and C. sonorense

Raw DNA sequence reads for C. pallidicaule USDA accession PI 478407 (BYU 1652; ^{Jarvis et al., 2017}), C. watsonii accession BYU 873 (Yavapai Co., Arizona), and C. sonorense Benet-Pierce & M.G. Simpson accession BYU 17220 (Santa Cruz Co., Arizona) were generated by the Novogene Corporation (Chula Vista, California, USA) using an Illumina (San Diego, California, USA) high-fidelity, short-read platform. Raw reads were then trimmed using Trimmomatic (^{Bolger et al., 2014}), removing low-quality reads and Illumina adapters. Reads from each accession were aligned to the quinoa QQ74 published reference genome (^{Jarvis et al., 2017}) using BWA-MEM (^{Li, 2013}), and SAM files were then converted to BAM format, sorted and indexed using SAMtools v1.9 (^{Li et al., 2009}). The GATK CollectAlignmentSummaryMetrics subprogram was used to generate mapping statistics (^{McKenna et al., 2010}).

Collection trips to Texas, New England, and Oklahoma in 2018

The trip to southern Texas in April resulted in the collection of seeds from 23 populations of C. berlandieri var. berlandieri; eight sympatric populations of putative diploid, broad-leaved C. albescens, which at one location (BYU 1816) was intermixed with a third narrow-leaved species that might have been C. pratericola; and six populations of the Gulf Coastal ecotype C. berlandieri var. boscianum (Table 1, Fig. 1a). The var. berlandieri populations all emitted a fetid trimethylamine odor and were mostly found on highly disturbed roadsides, including areas where seasonally dry riverbeds intersected the highway. The var. boscianum populations also smelled strongly of trimethylamine, were restricted to sandy barrier islands and tidal estuary margins, tended towards later maturity, and were more glabrous than var. berlandieri. Both ecotypes included populations of plants having semi-compact to compact inflorescences, with and without extensive stem branching. As shown in Fig. 2, var. boscianum included populations with significantly larger fruits, which approached 1.8 mm in diameter.

Figure 2 (a) Comparison of fruits of C. berlandieri vars. berlandieri and boscianum from southern Texas. In the columns, from left to right: var. berlandieri populations BYU 1814, BYU 1818, BYU 1822, and BYU 1833; var. boscianum populations BYU 1821, BYU 1825, BYU 1828, and BYU 1829. (b) Typical habitat of C. berlandieri var. berlandieri (BYU 1806) on cracked alkaline clay with Helianthus spp. (c) Typical habitat of C. berlandieri var. boscianum (BYU 1827) in sandy substrate on Mustang Island. 

The collection trip to the New England coast in September yielded seeds from nine populations of C. berlandieri var. macrocalycium (Table 1 and Figs. 1b and 3). This was a fairly common species in disturbed beach sand, especially among decaying seaweed at the high-tide line. Plants displayed a consistent phenotype with excessive branching, a yellow-green and glabrous appearance, and fruits to 2 mm in diameter in small, widely spaced glomerules and fairly uniform in maturity. At the Chatham Estuary of Cape Cod, populations extended a few meters into adjacent deciduous forest, where they were noticeably shorter in the shade. Notably, at least one population on the New Hampshire coast was being attacked by a stem-weakening, boring maggot (Fig. 3c).

Figure 3 (a) Fruits of six C. berlandieri var. macrocalycium populations from New England. In the columns, from left to right: BYU 1856, BYU 1857, BYU 1859, BYU 1860, BYU 1863, and BYU 1864. (b) Var. macrocalycium growing typically along the high-tide line among seaweed detritus at Sandwich Beach, Cape Cod, Massachusetts (BYU 1856). (c) Unidentified stem-boring maggot infecting var. macrocalycium plants at Odiorne Point State Park, New Hampshire. 

The collection trip to Oklahoma in October (Table 1 and Fig. 1c) resulted in eight collections of C. berlandieri var. sinuatum, four of var. zschackei, and eight of a hitherto unidentified variety with characteristics similar to those of vars. zschackei (yellow area at the stylar base) and berlandieri (strong trimethylamine odor and more conical fruit side profile). Figure 4 compares fruits from three populations of each of these ecotypes. Variety sinuatum was more common in northeastern Oklahoma, with var. zschackei predominant in the southeast and the unnamed variety more common in the western part of the state. At all sites, the species was strongly associated with disturbed, sandy soils and was especially common along roadsides that intersected with arroyos and seasonally flooded riverbanks.

Figure 4 (a) Comparison of fruits of C. berlandieri from Oklahoma. In the columns, from left to right: var. sinuatum collections BYU 1878, BYU 1879, and BYU 1880; var. unknown collections BYU 1883, BYU 1886, and BYU 1887; and var. zschackei collections BYU 1899, BYU 18101, and BYU 18102. (b) Fruits of Chenopodium subsect. Leiosperma diploids collected in Oklahoma. Clockwise from middle left: BYU 1881 (Alfalfa Co.); BYU 1888 (Woods Co.); BYU 1892 (Woods Co.); BYU 18100 (Dewey Co.); and BYU 18108 (Love Co.). 

In addition to the C. berlandieri accessions, we collected seeds from 12 populations of unidentified narrow-leaved species bearing utriculate fruits. We are currently working to identify these putative diploids via fruit morphology and DNA analyses.

Recovery of significant breeding traits from C. berlandieri x C. quinoa hybrids

We herein provide some examples that demonstrate the value of ecologically variable pitseed goosefoot as a quinoa improvement resource. These examples are from intertaxa populations derived via either spontaneous field crosses or intentional greenhouse crosses (Fig. 5).

Figure 5 Characteristics of plants derived from hybridization between C. berlandieri and C. quinoa. (a) Plot of selected F_2:5 plants (O108-1-11-17-1 parent) derived from a cross between Salares quinoa cv. ‘Ollague’ and BYU 14108 growing at the University of California Hansen Agricultural Experiment Station near Los Angeles, California, in June 2018. BYU 14108 (var. sinuatum) was collected in October 2014 in the torrid upper Sonoran oak-grassland savanna zone near the Chiricahua Mountains in southern Arizona. (b) Variation in fruit size and color in the ‘Real-1’ (top row, left) and BYU 937 (top row, middle) parents, F₁ plants (top row, right) and fifteen F₂ plants of population R1R-2. BYU 937 (var. boscianum) was a single plant collected along the shore of the Galveston Bay estuary, Texas, in August 2009. (c) Fruits from one early-maturing F₂ derived from a quinoa cv. ‘Brightest Brilliant Rainbow’ x var. macrocalycium hybrid showing transgressive segregation for large seed size. The F₁ parent was identified in a row of Brightest Brilliant Rainbow growing at the Woodman Research Farm at the University of New Hampshire Agricultural Experiment Station in September 2018, and seeds from that plant were kindly provided by UNH Professor Tom Davis; 17129 = var. macrocalycium parent; BBR = Brightest Brilliant Rainbow. 

Figure 5a shows vigorous and relatively uniform F_2:5 plants (O108-1-11-17-1 parent) derived from a cross between ‘Ollague’ and BYU 14108, a var. sinuatum accession collected in October 2014 in a torrid oak-grassland steppe in Cochise Co., Arizona. A single black-seeded hybrid plant was identified in the greenhouse after growing seeds harvested from a heat-stressed ‘Ollague’ plant exposed to BYU 14108 pollen. Among the 178 F₂ plants showing widely variable morphology in terms of plant architecture, seed size, and seed coloration, 169 (94.9%) produced at least one seed, and all but a few produced many more than five seeds. Only two of the plants, however, produced seeds that were not black, one of which (O108-1-11) bore white seeds that were similar in size to those of ‘Ollague’.

Figure 5b shows representative fruit size and color variation in the parent, F₁, and F₂ plants (population R1R-2) derived from a cross between quinoa cv. ‘Real-1’ and BYU 937, a var. boscianum accession collected in August 2009 at the high-tide line along the sandy shore of Galveston Bay, Texas. Multiple black-seeded, foul-smelling hybrids were recovered from seeds taken off a heat-stressed Real-1 plant that had been exposed to BYU 937 pollen in the greenhouse, one of which produced population R1R-2. The hybrid bore seeds that were similar in size to those of ‘Real-1’, while several F₂ plants produced seeds that exceeded the size of those of the quinoa parent, despite the fact that the var. boscianum parent produces seeds that are only slightly larger than 1 mm in diameter (Fig. 5b). This transgressive segregation pattern suggests that BYU 937 harbors at least one complementary allele for large seed size.

During the New England collection trip in September 2018, inspection of a row of ‘Brightest Brilliant Rainbow’ quinoa at the University of New Hampshire Experimental Farm in Durham revealed the presence of a putative F₁ plant that was very large and bushy with some phenotypic characteristics reminiscent of var. macrocalycium. Approximately 60 large, black seeds from this plant were kindly provided by Tom Davis and subsequently planted in the greenhouse at BYU, resulting in plants with a wide array of whole-plant and panicle morphologies, from compact to exceptionally lax. Figure 5c presents seeds from one early-maturing, putative F₂ showing transgressive segregation for large seed size. We previously noted in an ‘Ingapirca’ x var. macrocalycium cross that segregants in the F₂ and subsequent generations were universally heat susceptible, produced less than 50% fertile progeny, and produced no individuals with large seeds that did not also have highly lax panicles (personal observations).

Comparison of genomic DNA sequences of C. watsonii, C. sonorense, and C. pallidicaule with the quinoa QQ74 reference genome

Table 2 presents a summary of trimmed short-read sequence mapping of three A-genome diploids – wild Watson's stinking goosefoot, wild Sonoran goosefoot and cultivated cañahua – to the lowland QQ74 quinoa reference genome reported in ^{Jarvis et al. (2017)}. Although all three species are closely related to quinoa and belong to the A-genome diploid group, sequencing reads of Watson's stinking goosefoot showed the highest mapping percentage and the lowest mismatch rate, whereas cañahua showed the lowest mapping percentage and the highest mismatch rate.