Introduction

Seed banking is generally the most effective and efficient option for preserving plant germplasm ex situ, but it is not applicable to all species. Exceptional species, such as those with seeds that are sensitive to the drying required in seed banking procedures, and those with few or no seeds available for banking, require alternative methods (Pence 2013). One such method, shoot tip cryopreservation, has been used on a large scale for preserving genetic diversity of several clonally propagated crop species (Reed 2001), including staples such as Solanum tuberosum, Ipomoea batatas, and Musa spp., ornamental crops, fruit tree crops, and others (Fukai et al. 1991; Reed et al. 1998; Panis et al. 2005; Jenderek et al. 2013; Martín et al. 2013; Park and Kim 2016; Vollmer et al. 2016). Cryopreservation has been used less frequently for the conservation of wild plant germplasm (Touchell et al. 1992; Wilkinson et al. 2003; Pence et al. 2009; Pence 2015), but there has been growing interest in using this tool to preserve plant biodiversity (Engelmann 2011; Pence 2014).

Hedeoma todsenii R.S.Irving (Todsen’s pennyroyal) is a federally endangered plant that is an exceptional species because of the lack of sufficient seed production for banking (U.S. Fish and Wildlife Service 2001). Protocols were developed at the Center for Conservation and Research of Endangered Wildlife (CREW) (Cincinnati Zoo & Botanical Garden, Cincinnati, OH) by Pence et al. (2009) for the cryopreservation of shoot tips from in vitro cultures of this species, using both the encapsulation dehydration (ED) (Fabre and Dereuddre 1990) and the encapsulation vitrification (EV) procedures (Hirai and Sakai 1999), and samples were cryopreserved and stored intermittently over time in the Frozen Garden of CREW’s CryoBioBank™ starting in 2001. While many species have been shown to survive exposure to liquid nitrogen (LN), there is much less empirical evidence as to their ability to survive after years of LN storage, and most of the datasets available come from cultivated species (e.g., Mix-Wagner et al. 2003; Caswell and Kartha 2009). As part of a recent study to evaluate survival of a wide variety of samples in CREW’s Frozen Garden, samples of H. todsenii shoot tips were removed and evaluated for survival after 7–13 yr of storage, to examine the effectiveness of shoot tip cryopreservation as a long-term ex situ conservation tool for this species. More recently, shoot tips of this species were also cryopreserved with a second encapsulation vitrification method (EV-2) (Matsumoto et al. 1995), and with droplet vitrification (DV) (Panis et al. 2005), and the pre-storage survival of these samples were compared in this study with pre-storage survival using the ED and the EV protocols.

In addition to the need for survival datasets, there has also been little investigation into the effects of long-term cryopreservation on the genetic composition of stored samples. Specifically, the samples that are initially banked are assumed to be genetically identical to the samples that survive cryopreservation. To confirm genetic fidelity of cryopreserved samples, a combination of microsatellite and sequence-related amplified polymorphism (SRAP) genetic markers was used to compare the genetic composition of a subset of genotypes from different accessions banked at separate times. Microsatellites were used to assess genetic changes that may have occurred in mutation-prone repeat motifs of non-coding regions, or in regions that underlie where the primer(s) might anneal. SRAP markers preferentially target open reading frames and were used to expand the genome region sampled by microsatellites, and to verify the microsatellite results.

Materials and Methods

Plant material and growth conditions

Shoot tips for cryopreservation protocols were isolated from in vitro shoot propagating cultures of H. todsenii R.S. Irving (Lamiaceae) that were initiated as previously reported (Pence et al. 2009; Fig. 1). Unless otherwise indicated, all chemicals were from Sigma-Aldrich® Chemical Company, St. Louis, MO. All media were adjusted to pH 5.6 ± 0.1 with 1 M KOH and sterilized by autoclaving at 121°C, 138 kPa, for 15 min. All chemicals for culture media were added before autoclaving except abscisic acid, which was filter sterilized by passing through a 0.2-μm sterile filter (Thermo Scientific™ Nalgene™ Syringe Filter, Thermo Fisher Scientific®, Waltham, MA) and was added to the medium after autoclaving. For cryopreservation procedures, all PVS2 solutions were filter sterilized through a 0.2-μm sterile filter (Thermo Scientific™ Nalgene™ Rapid-Flow™ 150 mm Filter Unit, Thermo Fisher Scientific®), as were all other solutions for the pre-storage DV procedure, and the rinsing solution for some of the pre-storage EV procedures and all of the post-storage EV recoveries. All other solutions for the pre-storage ED, EV, and EV-2 procedures were sterilized by autoclaving under the conditions given above.

Fig. 1.
figure 1

A schematic of the initiation of the H. todsenii genotypes cryopreserved in this study. After initiation, cultures were subcultured at 2–3-mo intervals, with 5–10 culture tubes maintained per genotype. At times t and t + 1, shoot tips were harvested from a culture tube and cryopreserved, with 5–10 shoot tips in a single cryovial.

Full size image

Cultures were maintained on a medium of Murashige and Skoog (1962) salts and organic compounds (MS medium) (PhytoTechnology Laboratories®, Inc., Monmouth Junction, NJ), with 2.22 μM benzylaminopurine (BAP), 0.27 μM naphthalene acetic acid (NAA), 3% (w/v) sucrose, and 0.25% (w/v) gellan gum (Gelzan™, Caisson Laboratories, Smithfield, UT) (=maintenance medium, MM), in 25 × 150 borosilicate glass culture tubes, 15 mL of medium per tube, with colorless Magenta™ caps, at 26°C, under cool white fluorescent lights (Philips F32T8, Becker Electric Supply, Cincinnati, OH) and a 16-h light and 8-h dark cycle, approximately 60 μmol m−2 s−1 PPFD, and subcultured at approximately 2–3-mo intervals. Genotypes banked at different times were maintained in active culture throughout that period to provide a source of shoot tips.

Cryopreservation methods

Several cryopreservation protocols have been used to prepare H. todsenii shoot tips for long-term cryostorage (see below). For each banking, a small sample of shoot tips (10–20) was tested after short-term (1–24 h) exposure to liquid nitrogen (LN) to determine the pre-storage survival of the sample. Samples stored long-term were maintained submerged in LN in a Thermolyne® Dewars Locator® 4 (Thermo Fisher Scientific®, Inc., Pittsburgh, PA), 2000 vial capacity tank for 7–13 yr. For this study, a total of 187 vials of 10–25 shoot tips each were removed after long-term storage, representing 19 genotypes of H. todsenii (defined as a genetically distinct entity) distributed across 28 dates of banking. Each genotype banked at a specific date was considered an accession, and there were generally replicate vials banked per accession. Growth of these tissues was monitored weekly, with final survival data recorded 4 wk after recovery plating (Fig. 2). Post-storage survival was defined as the percentage of shoot tips that remained green and showed some shoot outgrowth by 4 wk. This was used as a marker for survival during the critical first few weeks and was consistent with the data taken at 4 wk to measure pre-storage survival in samples cryopreserved 7–13 yr earlier. In order to compare survival with regrowth, the growth of a subset of shoots exposed to LN for 1 h after cryoprotection using the DV method was evaluated 2–3 mo after rewarming. After long-term storage of shoot tips cryopreserved using ED and EV, some shoot tips were also grown into shoot cultures to provide material for genetic analysis (Fig. 2).

Fig. 2.
figure 2

A schematic of the procedure for recovery of each H. todsenii genotype in this study. When a replicate vial from an accession was removed from cryostorage and rewarmed, the shoot tips were placed on a plate of recovery medium. Shoot tips that successfully grew were subcultured into culture tubes. Tubes of cultures generated from the same plate were considered identical, and one was randomly chosen as a representative of the cryovial for genetic analysis. Time was variable (see text for details).

Full size image

Shoot tip preparation

For the encapsulation vitrification (EV), droplet vitrification (DV), and encapsulation dehydration (ED) procedures, tissues were subcultured into tubes of MM with 0.8% (w/v) agar (Sigma A1296) rather than gel for 4–6 wk, to prepare for cryopreservation. For the EV-2 procedure, most cultures were grown on MM with 0.8% (w/v) agar in large, 15 × 150-mm disposable Petri dishes (Falcon® Corning®, Tewskbury, MA) for 2.5–3 wk. When grown in tubes, shoots of H. todsenii produced very tiny leaves, which often showed a curling response that is not observed on the plant in situ Fig. 3 A. Similar growth with less curling of the leaves occurred in plates Fig. 3 B. In the case of both sources, tips of approximately 1 mm in length were isolated Fig. 3 A, inset for cryopreservation. For the ED, EV, and DV protocols, isolated shoot tips were pre-cultured on MS medium with 0.44 μM BAP, 0.05 μM NAA, 10 μM abscisic acid (ABA), 0.3 M mannitol, and 0.25% (w/v) gel in 60 × 15-mm disposable Petri plates, 15 mL of medium per plate, and incubated at the same temperature and light cycle as the stock cultures, but at a lower light level of 5–10 μmol m−2 s−1 PPFD, for 2 d. For the EV-2 procedure, shoot tips were prepared using other solutions as described below. Containers used for LN storage for all of the protocols were 2 mL polypropylene cryovials (Corning®, Corning, NY). For all encapsulation procedures, 3% (w/v) alginic acid was used with 100 mM CaCl2 to form the beads (Fabre and Dereuddre 1990; Hirai and Sakai 1999), while a rinsing solution of MS plus 1.2 M sucrose was used in the three vitrification procedures.

Fig. 3.
figure 3

(A) Shoot culture of H. todsenii grown in culture tubes on MM plus agar, used as a source of shoot tips for cryopreservation by EV, ED, and DV protocols, and (inset) examples of isolated shoot tips. Bar = 5 mm (inset 1 mm). (B) Shoot culture used as a source of shoot tips for the EV-2 protocol. Bar = 20 mm.

Full size image

Encapsulation dehydration (ED)

The procedure of Fabre and Dereuddre (1990) was followed for encapsulating shoot tips in beads of alginic acid and incubating them overnight with gentle swirling in a 0.75 M sucrose MS solution. The beads were then dried for 4 h under the air flow of a laminar flow hood, transferred to cryovials with 10–25 beads per vial, and cooled rapidly by immersing the cryovials in LN. After short- or long-term storage, the vials were rewarmed at room temperature (24°C) for 15 min, and the beads were then transferred to plates of recovery medium consisting of MS medium with 3% (w/v) sucrose, 0.44 μM BAP, 0.05 μM NAA, and 0.8% (w/v) agar in 60 × 15 mm plastic Petri plates.

Encapsulation vitrification (EV)

The procedure of Hirai and Sakai (1999) was followed for encapsulating the shoot tips in beads of alginic acid. Beads were then transferred to PVS2 solution (Sakai et al. 1990), incubated on ice for 2 h, and transferred to cryovials containing 1 mL of cold PVS2 with 10 beads per vial. The vials were immersed in LN for rapid cooling. Warming was completed after short- or long-term storage by immersing the cryovial in a water bath of approximately 45°C with gentle agitation for 2 min, after which the PVS2 solution was quickly removed and replaced with rinsing solution. After 10 min, the solution was removed and replaced with another 1 mL of rinsing solution for 10 min, and the beads were then transferred to the same recovery medium as for ED.

Encapsulation vitrification-2 (EV-2)

A second procedure, based on that of Matsumoto et al. (1995), was used for shoot tips cryopreserved from 2010 to 2011. After excision, the shoot tips were encapsulated in alginate beads and transferred to a pre-incubation medium (MS plus 1 g L−1 casamino acids) for 1 d, followed by a pre-culture medium (pre-incubation medium plus 0.3 M sucrose) for 16 h. Next, the beads were transferred to an osmoprotection solution (MS plus 2 M glycerol and 1.6 M sucrose) for 4 h, followed by PVS2 at room temperature (24°C) for 1 h. The beads were then transferred to a cryovial with 1 mL of PVS2, and the vial was rapidly immersed in LN. After short-term storage, the vials were rewarmed in a water bath of 40–45°C for 2 min. PVS2 was immediately replaced with rinsing solution, and after 10 min the solution was removed and replaced with another 1 mL of rinsing solution for 10 min. The beads were then moved to a recovery medium of MS medium with 3% (w/v) sucrose, 4.44 μM BAP, and 0.25% (w/v) gel.

Droplet vitrification (DV)

After excision, shoot tips were incubated for 2 d on the pre-culture medium used for the ED and EV procedures. They were then cryopreserved using the procedure of Panis et al. (2005). The shoot tips were incubated in a loading solution (MS plus2 M glycerol, 0.4 M sucrose) at room temperature (24°C) for 20 min, followed by cold PVS2 for 20 min. The shoot tips were then moved to a small strip of aluminum foil (Great Value™ Aluminum Foil, Distributed by Wal-Mart Stores, Inc., Bentonville, AR, Product of Armenia), hand-cut to approximately 0.8 × 3.0 cm, with a small amount of PVS2 and were rapidly cooled by immersion of the foil and tips directly into LN. The foil strip, including the cooled tips, was placed in a cryovial, pre-chilled in LN, but not containing LN. After short-term storage, the tips were rewarmed by rapidly opening the vial and removing the foil strip plus tips and immersing it into rinsing solution at room temperature (24°C). After 15 min, the tips were transferred to the same recovery medium used for the ED and EV procedures.

Genetic analysis

A subset of four H. todsenii genotypes that had each been cryopreserved at separate time points was used to assess genetic fidelity (Figs. 1 and 2). Two genotypes were banked at three time points each; one genotype was banked at two time points, and one genotype was banked at one time point (Table 1). Because cryopreservation of these samples was first performed 7–8 yr ago, before genetic analysis in these types of studies was as commonly used, genetic studies were not conducted on genotypes prior to cryopreservation. DNA extractions were performed on fresh leaf tissue isolated from cultures grown from one recovered shoot tip per vial (Fig. 2) using a modified cetyltrimethylammonium bromide (CTAB) protocol (Doyle and Doyle 1987). One shoot tip per vial was used as a proxy for the entire vial, as individual shoot tips within a vial were replicates of each other, and presumed to be identical (Fig. 1). Surviving plants regrown from shoot tips within a vial were pooled for continued maintenance, and thus, one representative per vial was sampled.

Table 1. Genetic analysis of replicate vials of accessions (unique banking times) representing different genotypes of Hedeoma todsenii, using microsatellites (MS) as percentage band loss of seven microsatellite loci and SRAPs as percentage band loss or change (polymorphism)
Full size table

Microsatellite analysis was performed using seven primers (Integrated DNA Technologies, San Jose, CA) identified from two other Lamiaceae taxa, Conradina spp. (Primers BR50-2, CSRC1, CSRF, CSR30-4, GLA-H, GLA30-4; Edwards et al. 2008), and Suzukia shikikunensis (Primer SU09; Hung et al. 2008). Microsatellite primers were multiplexed according to Culley et al. (2013), and alleles (hereafter referred to as “bands”) were visualized using fragment analysis on data generated from the ABI 3730xl sequencer (Thermo Fisher Scientific®, Florence, KY). All samples were analyzed four times to confirm the band calls, and a consensus sequence was created from replicate runs for each sample to minimize error rates.

SRAP analysis was performed using eight primer combinations, which produced 22 scoreable fragments visualized as bands on 2% (w/v) agarose gels, according to Li and Quiros (2001). All samples were run in triplicate to confirm the SRAP banding patterns. For each genotype, the genetic fingerprints generated by the two types of molecular markers were compared between cryopreserved accessions (unique banking times) to assess the total number of band matches (Fig. 2). Multilocus matches for both marker types were calculated as the percentage of matching bands within a genotype (across replicates). For microsatellites, band loss was determined by the absence of a band in all replicates of a genotype. For SRAP markers, band polymorphism was determined by the number of bands gained or lost in a single accession, compared to the total number of bands scored within a genotype.

Data analysis

Statistical analyses were made using R version 3.3.2 (https://cran.r-project.org/bin/windows/base/old/3.3.2/). To test long-term survival, a survival analysis was performed to determine the relative risk contributed by genotype, banking procedure, and time banked using the Cox proportional hazards model (Cox 1972). The covariates did not violate the proportional hazards assumption. The survival analysis was performed in R using the survival package (Therneau 2015). Because the survival analysis analyzed the survival of individual shoot tips instead of averaged survival per vial, there is some degree of pseudo-replication confounding this analysis. This is unavoidable when calculating time-to-death measurements for shoot tips due to the nature of shoot tip cryopreservation, but it should be noted when interpreting survival analysis results. To compare the outcomes of the different cryopreservation methods, non-parametric tests were used due to zero inflation and non-normality in the data. Differences among the four cryopreservation methods were analyzed with a Kruskal–Wallis test (Kruskal and Wallis 1952), and post hoc analyses were performed using the dunn.test package (Dinno 2017). To account for any survival differences due to genotype, a generalized linear mixed effects model was used to assess the relationship between survival and banking procedure, using genotype as a random effect. Models were generated using the lme4 package (Bates et al. 2015), and post hoc analyses were performed using the multcomp package (Hothorn et al. 2008). Graphing was completed using JMP® 10.0.0 2012 (SAS Institute, Inc., Cary, NC).

Results

Long-term survival

The average post-storage percent survival of H. todsenii shoot tips cryopreserved with either the ED or the EV procedures and stored from 7 to 13 yr was not significantly different from the pre-storage percent survival with those methods (Fig. 4), but the variability between accessions within genotypes was high. In addition, a linear model indicated that there was no significant relationship between an accession’s pre- and post-storage survival values (R 2 = 0.02, F 1,32 = 1.8, p = 0.19) Fig. 5 A. There was also no relationship between the time in storage and post-storage viability compared with pre-storage viability Fig. 5 B. Whereas most samples were evaluated only at 4 wk, a subset of 18 H. todsenii genotypes cryoprotected using DV and exposed to LN for 1 h was also evaluated 2–3 mo after rewarming. These samples showed a total average survival rate of 74%, but only 46% displayed shoot growth greater than 1.0 cm (up to 15 cm) 2–3 mo after rewarming (data not shown).

Fig. 4.
figure 4

Average pre-storage survival (LN exposure of 1 h) of shoot tips cryopreserved using the ED, EV, EV-2, and DV protocols, and post-storage survival (7–13 yr in LN) of shoot tips cryopreserved using ED and EV. Standard error bars, the number of genotypes represented, and then number of samples analyzed (in parentheses) are given for each column. N numbers were larger for post-storage averages, as replicate samples were removed, compared with one sample for each accession analyzed for pre-storage survival. Data for all samples (matched and unmatched) are presented.

Full size image
Fig. 5.
figure 5

(A) Relationship of pre- and post-storage survival of H. todsenii samples for which both pre- and post-storage survival data were available. Samples were stored for 7–13 yr in liquid nitrogen using the EV (red square) and ED (blue diamond) methods. Dashed line indicates projected results for no change in viability during storage. (B) Relationship of time in storage and the change in survival (pre-storage viability–post-storage viability) for samples with both pre- and post-storage viability data, stored with EV (red square) and ED (blue diamond) methods

Full size image

The survival analysis used to examine the effects of genotype, banking procedure, and the time banked revealed some significant genotype effects on survival. In particular, the risk of death in genotypes 6 and 12 was 76% lower than the remaining genotypes (Fig. 6, p < 0.01). Controlling genotype as a random variable produced a better model, and there was no significant effect of the banking procedure or time banked on the survival of samples. This indicated that certain genotypes may be more naturally adapted to survive the stress of tissue culture and cryostorage, regardless of the cryopreservation procedure or time in cryostorage.

Fig. 6.
figure 6

Survival probability after removal from long-term cryostorage. Probability of survival after 1, 2, 3, and 4 wk post removal from cryostorage is displayed for the genotypes used in this study (each color represents a different genotype)

Full size image

Cryopreservation methods

When all four methods used to bank H. todsenii were compared, pre-storage survival rates using the ED, EV, DV, and EV-2 procedures ranged from 24.2% (EV-2) to 72.2% (DV) (Fig. 4). The Kruskal–Wallis test revealed significant differences in survival based on the banking procedure (χ 2 = 47.91, p < 0.001). A post hoc test using Dunn’s test with Benjamini–Hochberg correction showed that average pre-storage survival of EV-2 samples (24.2%) was significantly lower than EV samples (37.9%, z = 2.38, p = 0.03), and DV samples (72.2%, z = 6.87, p < 0.001). In addition, the average pre-storage survival rate of DV samples (72.2%), was also significantly higher than ED (31.0%, z = 3.42, p < 0.001) and EV samples (37.9%, z = 3.98, p < 0.01) (Fig. 4). Post hoc analysis of the generalized linear mixed effects model that incorporated genotype variation revealed that only the DV method was significantly higher than the other methods once genotype was controlled, suggesting that the low survival initially seen in EV-2 samples is due more to genotype effects than the procedure itself (EV-2 vs. EV, z = − 0.18, p = 0.998). Of the genotypes banked using the EV-2 procedure, 39% were also banked using the DV procedure, and the remaining 61% were only banked using the EV-2 procedure.

Genetic analysis

For the microsatellite analysis, seven loci revealed an average multilocus match (average number of matching bands between samples) of 89.71% among the four genotypes banked in different accessions or replicate vials (Table 1). Genotype 2 had the highest multilocus matches (94.64%) among replicates. Genotype 7 had 87.36% multilocus matches among replicates, genotype 11 had 89.86%, and genotype 12 had 86.96% multilocus matches among replicates.

For the SRAP analysis, 22 scoreable bands revealed a similar level of average multilocus matches (89.96%) as the microsatellite markers. The lowest percentage of multilocus matches across replicates was seen in genotype 2 at 79.55%, and the highest value was seen in genotype 11 at 95.45%. Genotypes 7 and 12 both had 92.42% multilocus matches across replicates.

Discussion

This study demonstrated that good survival rates can be obtained from shoot tips of H. todsenii after up to 13 yr in cryostorage, which supports shoot tip cryopreservation as an effective, alternative tool for the conservation of H. todsenii, because seed banking is not possible with this species. While several cryoprotective protocols have been shown to result in effective survival of shoot tips of many species through short-term exposure to LN (e.g., 1 h, 1 d) (Gonzalez-Arnao and Engelmann 2006; Sakai et al. 2008), there have been relatively few datasets documenting longer-term survival. One of the longest storage tests was for Pisum sativum and Fragaria x ananassa shoot tips after 28 yr of cryostorage using the two-step, slow freezing method (Caswell and Kartha 2009). In these cases, post-storage survival was equal to pre-storage survival for Fragaria, but there was a large decline in viability for P. sativum. In shorter-term studies, shoot tips of Wasabia japonica showed no loss in viability after 10 yr in cryostorage at − 150°C, using vitrification and initial cooling in LN (Matsumoto et al. 2013), while Dianthus hybridus showed no loss of viability over 4 yr of storage after banking using a slow freezing method (Fukai et al. 1991). Good survival has also been reported after shorter storage times, including 2 yr with Carica papaya using vitrification (Wang et al. 2005), and 1 yr for Cosmos atrosanguineus, an endangered species, using encapsulation dehydration (Wilkinson et al. 2003). As with the latter species, these results with the endangered H. todsenii support the use of cryopreservation as a tool for conservation, although results may vary with species, which requires refinement of protocols for some taxa.

In this study, survival was measured as the percentage of shoot tips that remained green and showed some regrowth by 4 wk, and this was used as a marker to study the effectiveness of the cryopreservation protocols used and the persistence of the samples through long-term storage. Studies with seed germination and with cryopreserved shoot tip regrowth have indicated that recovery of plants from such systems is generally lower than the initial seedling germination or shoot tip survival (Wilkinson et al. 2003; Sharaf et al. 2012; Park and Kim 2016; Poisson et al. 2016; Ballesteros and Pence 2017), and this must be considered to assess the ability of cryopreserved samples to adequately support any future needs for species recovery. However, the collection of growth data from cryopreserved shoot tips is not straightforward, as growth may occur at different rates, making the consistent timing of observations difficult. Additionally, even when significant shoot production does occur, those shoots may or may not go on to initiate a sustainable, shoot-producing culture, as has been observed with Cycladenia humilis var. jonesii and other species (Pence et al. 2017)). This may reflect an incomplete knowledge of optimal growing conditions for these species, which may change during the time the tissues are maintained in cryostorage. Thus, just as radicle protrusion is an accepted marker of seed viability (Walsh et al. 2003; Godefroid et al. 2010; Mira et al. 2011; Orrù et al. 2012), even though it has been shown that further growth of the seedlings may not always give rise to healthy plants (Popova et al. 2013; Ballesteros and Pence 2017; International Seed Testing Association 2017), early shoot outgrowth from cryopreserved shoot tips can be considered a sign of viability and potential for growth. However, the importance of growth in addition to survival as a measure of viability from cryopreserved shoot tips is increasingly being acknowledged (Wilkinson et al. 2003; Sharaf et al. 2012; Park and Kim 2016; Poisson et al. 2016), and will require further study in H. todsenii.

The initial survival levels using the ED and EV protocols with H. todsenii were lower than the 40% that has been proposed as the standard acceptable survival level for cryopreserved shoot tips (Reed et al. 1998; Jenderek et al. 2013). However, survival levels appeared to be adequate for recovering material from the clonal samples after long-term storage, and the survival averaged over multiple accessions after storage was similar to such values before storage. However, for individual accessions, pre-storage survival values did not predict post-storage survival, even though the average pre- and post-survival values were similar across genotypes and accessions. This may be due to a high level of variability between replicate samples within an accession, with some replicate vials showing no survival and some showing good survival, which has been observed by others (Caswell and Kartha 2009). Because of this variability, the use of one vial (generally10 shoot tips) to evaluate pre-storage viability, as was done with H. todsenii, may not be adequate to accurately represent the true viability of the accession. This indicates the importance of maximizing the number of replicate vials per species for cryopreservation work and using sample sizes that are as large as possible given the resources of the laboratory.

Cryostorage has been applied to other species of Lamiaceae, particularly species of Pycnanthemum spp. and Mentha spp. (Senula et al. 2007; Jenderek et al. 2013; González-Benito et al. 2016) using both ED and DV procedures. Survival of these species ranged from 40 to 100%, depending on the genotype, and was similar to the DV results obtained in the present study for H. todsenii but generally higher than ED or EV treatments on this species. Pre-treatments for the Pycnanthemum shoot tips included 2 wk of cold hardening in short day conditions before cryopreservation with ED. A similar pre-treatment had been shown to increase survival of Mentha shoot tips through DV cryopreservation, compared with pre-cryopreservation culture at constant temperature (Senula et al. 2007). Cold hardening has been shown to increase survival of shoot tips of other species through cryopreservation (Panta et al. 2015) and might improve survivals with H. todsenii through ED or EV. However, the improvements in H. todsenii survival observed with the DV method appear to overcome the limitations of ED for this species without the need for additional steps.

With H. todsenii, the DV treatment provided much higher survival rates than those achieved by the ED, EV, or EV-2 methods used, which were likely based on the more rapid cooling and rewarming that occurs in the DV method (Panis et al. 2005). This is an example of how technology is likely to improve during long-term storage, so when materials are rebanked, a different method may be used. Improvements in methods that provide higher survival rates should help mitigate against the small sample sizes that are inherent in work with endangered species cryopreservation, and which may account for the variability seen in survival between replicate vials of the same accessions.

In this study of H. todsenii, long-term cryopreserved samples showed high levels of genetic fidelity, although none of the four assayed genotypes exhibited 100% genetic fidelity between replicate vials or accessions for either type of molecular marker used in the analysis. On average, there was a 10.4% band loss observed between replicates within genotypes. This indicates that the stresses of cryopreservation and/or long-term storage, including the necessary step of tissue culture, may contribute to some loss of genetic fidelity. Furthermore, shoot tips were necessarily propagated through tissue culture following recovery from cryopreservation, making it difficult to determine whether genetic changes were due to cryopreservation stress or to somaclonal variation associated with tissue culture. Given that plants must be micropropagated following cryopreservation, it is difficult to determine exactly when any genetic changes may occur, as genotyping cannot occur until plants are large enough in tissue culture, to avoid unnecessary loss of samples in this federally endangered species.

In the literature, genetic changes following cryopreservation appear to be observed less often in microsatellites (Castillo et al. 2010; Hazubska-Przybył et al. 2013) than in non-specific markers like Amplified Fragment Length Polymorphisms (AFLPs), Random Amplification of Polymorphic DNA (RAPD), or SRAP markers (González-Benito et al. 2016). Castillo et al. (2010) found polymorphisms in Rubus grabowskii after 12 yr of cryostorage, but unlike the current study, polymorphisms were limited to their assayed AFLP markers and were not observed in their assayed microsatellites. Genetic change following cryopreservation has been reported using both types of markers, however, as seen by Zevallos et al. (2014), who found changes in microsatellite length in cryopreserved Solanum lycopersicum, and González-Benito et al. (2016), who found genetic changes in cryopreserved mint shoot tips using RAPD and AFLP markers. Given this variability in genetic stability, the genetic stability above 80% for nearly all samples of post-storage accessions of H. todsenii coupled with their lack of abnormal morphology lends further support to the use of cryopreservation as a conservation method for H. todsenii. However, plants recovered from cryopreservation should be monitored for longer periods of growth to determine if band loss has any morphological effects. Observed losses underscore the need for maximizing banking of replicate vials to preserve the genetic integrity of the genotype. To confirm this, it is also recommended that a DNA sample be taken from each genotype before banking to serve as a control for genetic fidelity. This practice would ensure that raw genetic material for individual genotypes would be available for future analyses regardless of advancements in technology, so the genetic consistency of cryopreserved samples could be tracked over time.

Conclusions

This report provides evidence that long-term storage through shoot tip cryopreservation is a viable option for conserving H. todsenii ex situ, with successful regeneration of material after up to 13 yr of storage. Furthermore, droplet vitrification can be an improved method for storage. Genetic fidelity between replicate samples was reasonably high, and when genetic changes were detected, it was usually within one or a few accessions of a genotype. However, genetic fidelity should be continuously monitored in this species due to the observed changes.