Bull kelp, Nereocystis luetkeana, is a species of brown algae within the Order Laminariales native to the West Coast of North America, with a range spanning from the Aleutian Islands in Alaska to Point Conception in Southern California. This species is of great importance both ecologically and economically: bull kelp canopies provide habitat for numerous coastal species (Rogers-Bennett and Catton 2019, Eger et al. 2023), and the species is increasingly being cultivated commercially as part of the nascent kelp aquaculture industry in Alaska (Alaska Department of Fish and Game 2024). Populations of bull kelp in its mid to southern range are being targeted for restoration due to intensive canopy losses as a result of the 2014–2016 marine heatwave (McPhearson et al. 2021, Tolimieri et al. 2023). Due to increased interest over the past decade in cultivating bull kelp for restoration and aquaculture, understanding the bull kelp lifecycle and variations within that lifecycle is of paramount importance.

Bull kelp possesses a heteromorphic lifecycle characteristic of the Order Laminariales (Fig. 1). The diploid, sporophyte stage of bull kelp (A) can grow up to 30 m tall, and at maturity will abscise sori (B) from the distal ends of its blades. Spores (C) can also be released from not yet abscised sori. The mature sori or newly released spores will sink to the benthos, where spores will settle, attach to suitable substrate, and eventually develop into microscopic, haploid male (D) and female (E) gametophytes. As female gametophytes exude eggs (F), they also release a hormone called lamoxirene that triggers the release of sperm (G) from male gametophytes (Maier et al. 2001). Fertilized eggs then begin to develop into new, diploid, embryonic sporophytes (H).

Within the past 20 years, a theory has emerged that bull kelp sori are capable of directly releasing embryonic sporophytes (I in Fig. 1) due to observations of sporophyte-like structures directly released from sorus tissue (Fig. 2). This theory was first postulated in a graduate thesis that utilized DAPI staining and image analysis to estimate the amount of DNA present in these non-typical sporophyte-like structures released from bull kelp sori (Kidder 2006). They concluded based on the amount of DNA present that diploid tissue was present in a structure that should be haploid, and thus they postulated that the unexpected structures were embryonic sporophytes. More recently, directly released structures were observed and documented to fluctuate seasonally in Alaska bull kelp populations (Ulaski and Konar 2021). Importantly, neither study measured the growth and development of these sporophyte-like structures over periods longer than 48 hours to track their developmental fate.

The timing of bull kelp microstage development under adequate light and nutrient conditions tends to be relatively consistent. Germination begins within a week of spore settlement, and by two weeks after release, gametophytes are present. By three weeks after release, female gametophytes begin to produce eggs. Between three and four weeks, fertilization takes place and early stage embryonic sporophytes begin to develop (Farrugia Drakard et al. 2023, Korabik et al. 2023, Weigel et al. 2023).

Here we seek to test the hypothesis that the sporophyte-like structures released from bull kelp sori, hereafter referred to as “directly released structures”, are collections of unreleased spores (sporangia) (J in Fig. 1) rather than embryonic sporophytes. We hypothesize that embryonic sporophytes should grow rapidly under ideal laboratory conditions into recognizable juvenile sporophytes, whereas sporangia or clusters of spores should undergo germination and subsequent gametophyte development.

In May 2024, sori were collected from the distal end of the longest mature blade of 20 adult bull kelp near Woody Island in Kodiak, Alaska. We also documented the condition of the adult bull kelp that sori were collected from. At collection, adult bull kelp were generally foliose with fresh blades, with more than 75% of the observed kelp possessing 15 blades or more. More than 50% of blades per individual possessed sori at a variety of stages of development. All sori collected appeared to be dark in color. Sori were immediately manually cleared of any fouling and rinsed with sterile seawater to remove any contaminants. They were then stored in a cooler between layers of damp paper towels overnight. The next morning, sori were soaked in a bath of UV-treated seawater and germanium dioxide solution (0.5 mL GeO₂ solution L⁻¹, as described in Shea and Chopin 2007) for 6 h, which produced a spore solution with a density of 177,500 spores per mL. After 6 h, 250 mL of spore solution was collected and passed through a 53-micron filter to remove any particles and mucilage.

We added 10 mL of filtered spore solution to a glass petri dish and examined the solution using a Leica DMLB fluorescence microscope (Wetzlar, Germany). When directly released structures were located, we retrieved each structure from the solution using a 10 μL pipette tip and transferred it into its own well in a 24-well plate. At least 24 directly released structures were isolated from the spore solution, but there were some instances when more than one directly released structure was retrieved at a time. Pipetting such small amounts of liquid was imprecise, and as such, normal spores may have been introduced to each well in addition to the directly released structures. The quantity of spores was likely negligible and did not affect our ability to later identify isolated directly released structures.

Well plates were covered and placed in a cold room (set at 12°C) under total darkness for 48 h to allow structure settlement and attachment to the bottom of the well plates. After 48 h, the well plates were placed under cool, white lights (30 μmol m⁻² s⁻¹) at a 12 : 12 diel cycle. Cultures were kept under these conditions for 3 weeks, during which medium in the well plates was changed weekly in order to replenish F/2 nutrients and add GeO₂ (0.5 mg L⁻¹ seawater) to limit diatom growth. Temperature was monitored via HOBO loggers placed in the cold room.

After 4 days, we photographed each well under 100× magnification using a Leica DMLB fluorescence microscope and Image Pro 11 Software (Media Cybernetics, Inc., Rockville, MD, USA) to document the number of directly released structures and other microscopic lifestages (collectively referred to as “microstages”) present in each well. We photographed every single microscopic structure seen in each well, whether it was a directly released structure (i.e., the hypothesized embryonic sporophyte), gametophyte, or other suspect cells (“Other” collectively refers to other spore clusters that did not take the same form as directly released individuals, miscellaneous refuse, and unidentifiable structures). This process was repeated at 7, 14, and 21 days. At the end of the growth period, the microscopic structures present in the photos were visually assessed and matched across the 4 time periods (4, 7, 14, and 21 days) to determine the development of the directly released structures over time (Supplementary File S1).

In order to compare the size of (1) released spores and cells within directly released structures, and (2) embryonic sporophytes and directly released structures, we used Image J (version 1.54g, National Institute of Mental Health, Bethesda, MD, USA) to assess the length of directly released structures, cells within the directly released structures, and spores upon initial spore release. We also measured the length and area of eggs, sporophytes, and directly released structures present after 21 days. Data was square-root transformed to achieve normality and homogeneity of variance. We ran t-tests (alpha = 0.05) with and without assumed homogeneity of variances as appropriate to compare the lengths of spores and cells within directly released structures at release and the lengths and areas of eggs, sporophytes, and directly released structures after 21 days. Even after square-root transformation, residuals for the area of eggs versus directly released structures were not normal, so we used a non-parametric Wilcoxon rank sum test.

Over the course of 3 weeks, we identified and tracked over 1,000 microstages. We saw different numbers of each microstage category at each time interval (Table 1). At 4 days after release, we observed 116 directly released structures, and 70 structures that fell into the category of “other”. At 7 days after release, we observed more directly released structures (n = 129) that likely needed more than 4 days to settle. We also began to see a small number of gametophytes (n = 4) already developing, as well as an increased number of “other” structures (n = 176). By 14 days after release, the number of directly released structures observed decreased (n = 21), while the number of gametophytes present increased (n = 277). Finally, on day 21, a similar number of directly released structures remained (n = 29), while the number of gametophytes more than tripled (n = 1,009) and embryonic sporophytes began to appear (n = 62). There were also 151 and 174 “other” structures observed on days 14 and 21, respectively.

Of these observed microstructures, we were successfully able to track 259 microstage structures across multiple time periods (Table 2). We tracked 67 directly released structures for more than one week (Table 3). None exhibited the growth expected of an embryonic sporophyte. As the directly released structures developed over the course of the 3 weeks, the size and shape of the structures remained the same, indicating that there was no growth or cell division occurring. Instead, the directly released structures seemed to lose cell structure and pigmentation, and in some cases nearly disappeared entirely (Fig. 3). In other cases, the directly released structures seemed to exhibit some cell growth on day 7 compared to day 4, but by day 14 hosted clusters of young gametophytes (Fig. 4). Young sporophyte growth was not evident until day 21, and even then, sporophytes had clear origins in a gametophyte.

T-tests revealed significant differences between the size of all compared structures (Table 4). At spore release, the size of spores was significantly larger (0.071 ± 0.003 mm) than the cells observed within the directly released structures (0.058 ± 0.008 mm; t = −10.807, df = 63.055, p < 0.001). At 21 days, when embryonic sporophytes originating from gametophytes were first observed, the size of sporophytes was significantly larger than the size of directly released structures, both in terms of length (t = 4.0913, df = 147, p < 0.001) (Fig. 5A) and area (t = 11.8, df = 90, p < 0.001) (Fig. 5B). Directly released structures were significantly longer than eggs released by female gametophytes (t = −7.6016, df = 151.42, p < 0.001) (Fig. 5A), but eggs had significantly larger areas than directly released structures (W = 1,512, p < 0.001) (Fig. 5B).

After culturing directly released structures for 3 weeks, we reject the hypothesis that directly released structures are embryonic sporophytes. Rather, we conclude that these directly released structures are sporangia. It is highly likely that the structures we observed are the same as those discussed in previous studies (Kidder 2006, Ulaski and Konar 2021), due to similar shape and cellular structure (Fig. 2). Directly released structures are also not large enough to be sporophytes, and are significantly smaller in area than eggs released by female gametophytes (Fig. 5).

The premise of direct embryonic sporophyte release posed by Kidder (2006) was based on previous observations of development of “haploid sporophytes” produced through the parthenogenetic development of an unfertilized egg (Kemp and Cole 1961). Kemp and Cole (1961) described these structures as stunted and only producing two to three nucleated cells after five months. This description and the supporting photos of parthenogenetic development seem to describe a very different structure from the mass of tens of cells originating from sorus tissue that was described in Kidder (2006) and Ulaski and Konar (2021). Instead, the development of the directly released structures examined in our study more closely followed the pattern of the sporangia described by Kemp and Cole (1961). Sporangia begin development in the sori of their parent plant and proceed through several rounds of mitotic division until they are comprised of 32 uninucleate protoplasts. These protoplasts then morph into zoospores that are eventually liberated via the rupture of the tip of the sporangium. At initial release, the putative sporangia that we observed match Kemp and Cole’s description of sporangia pre-rupture and spore liberation (Fig. 6).

Kidder (2006) proposed that the directly released structures were diploid based on estimated DNA content ratios; however, the evidence to support this conclusion is limited. Kidder’s method inferred ploidy by DAPI staining DNA and then using fluorescence microscopy to estimate the ratio of DNA-containing area to whole-cell area for the adult sporophyte (diploid, 2n), spore/gametophyte (haploid, n), and putative embryonic sporophyte structure. This derived DNA content ratio has multiple challenges: differences in whole-cell area across stages (e.g., the spore vs. protoplast) would alter this ratio calculation, as would any changes to DNA area. Unfortunately, only the ratio is provided in Kidder (2006) and the photomicrograph images are over-exposed, making it impossible to assess the fluorescence area or intensity. More definitive methods for evaluating ploidy would be flow cytometry using known standards for diploid or haploid cells, counting chromosomes, or even targeted genotyping with a modest (<10) number of genetic markers (Bohanec 2003, Pellicer and Leitch 2014). In order to conclusively propose an alternate life history (i.e., the presence of diploid embryonic sporophytes released directly from sorus tissue), verification of ploidy using these accepted methods would be necessary.

In addition to evaluation of the DNA content of these directly released structures, Kidder (2006) argues that when these structures were ruptured, spores did not “fall out” as would be expected of sporangia. This is hardly surprising, however, when considering the nature of the sporangial structure. Within a given sporangium, spores develop in a matrix of mucilage that is then exudated in large quantities during spore release (Walker and Bisalputra 1977). This mucilage is generally viscous and sticky, and N. luetkeana spores released under laboratory conditions are frequently observed to clump together (Fig. 7). Natural spore release in N. luetkeana involves sorus abscission, which has been described at the ultrastructural level as consisting of phases of necrosis and dissolution of specific cells and tissue layers, including a dissolution of the cuticle covering the sporangium (Amsler and Neushul 1989) and the release of mucilage (Walker and Bisalputra 1977). Taking all this into account, it is unsurprising that the rupturing of immature sporangia did not lead to spore release—these were likely held inside the sporangia due to the adhesive nature of the sporangial matrix. If diploid genetic material was indeed detected in these putative sporangia, it is likely a result of the fact that the sporangial sheath is composed of diploid parent material and the mucilage matrix likely also contains diploid genetic material.

More research is needed to fully investigate the cellular structure and function of these sporangia, as well as answer questions about why they are only seen in some releases and not others. One question that arises is whether the direct release of sporangia is related to sorus age and development. The temporal patterns described in Ulaski and Konar (2021) indicate that sporangia release is highest in spring and decreases as the growing season progresses (with an unexpected absence in June). In other words, kelp that only just reached the surface in April may require more time to mature, but by September, releases consist almost entirely of normal spores. The sori in our experiment may indeed have been underripe, which may possibly explain the presence of sporangia in our spore solution. Additionally, all of these experiments are based on “forced” release in the lab, which requires stressing sori to induce reproduction before death. We thus suggest that inducing reproduction in immature sori may force the release of these premature sporangia. Understanding the factors and timing guiding the release of sporangia, and release of spores from these sporangia, will be beneficial to better understand the nuances of the bull kelp life cycle.

Identification of these questionable biological structures presents an excellent opportunity to apply Occam’s Razor: in the absence of extensive genetic and physiological tests, it is more parsimonious that these structures represent a known phase of kelp development rather than a novel life history strategy. We posit that these structures are merely immature or mature sporangia and may indicate that the sampled sorus tissue was not yet completely ripe. Although our experiment did not conclusively determine the ploidy of the putative sporangia, a future study might use one or more accepted methods to identify the number of chromosomes in each of the relevant developmental stages. For example, standard methods using flow cytometry rely on the fluorescence values of the unknown sample compared to fluorescence of developmental stages with known ploidy levels, or for which ploidy is confirmed by complementary methods (e.g., chromosomal counts) (Tomaszewska et al. 2021). Alternatively, marker-based genotyping can distinguish between haploid and diploid life stages based on the presence of multiple DNA bases (alleles) at a single location in the genome (i.e., heterozygosity) (Gompert and Mock 2017). In the absence of data from one of these methods, but based on the development trajectories we observed by isolating and culturing the questionable structures, we have determined that these structures and those observed by Ulaski and Konar (2021) and Kidder (2006) are sporangia, not embryonic sporophytes.