T₀ transformants of the wheat and barley constructs were obtained in the Fielder and Golden Promise cultivars, respectively, and screened for copy number. We identified 17 independent events which contained the wheat construct, HS_NAM-B1, and eight independent events which contained the barley construct, HS_GFP (Additional file 2: Table S1). Fourteen independent wheat lines and four barley lines were taken forward for copy number analysis. From these lines, four wheat lines with varied copy number were taken forward for the majority of further analysis at the T₁ generation. Two single-copy barley lines were taken forward for analysis, one segregating population (00899-01-01) and one homozygous line containing a single insertion, and thus two copies, at the T₁ generation (00899-04-01).

To characterise the heat shock-inducible construct, we first tested whether the application of heat shock to positive transformants would be sufficient to induce expression of Cre recombinase. Based on previous work with the HvHSP17 promoter [13], we expected that 2 h at 38 °C would be able to induce expression of Cre. Using T₁ plants descended from the wheat 4-copy line 2020-2-1, we found that within 2 h of heat shock treatment to seedlings, Cre expression could already be detected (Additional file 2: Figure S4).

We investigated whether the expression of Cre was sufficient to cause successful excision of the reporter gene. In wheat seedlings before heat shock treatment, PCR of extracted genomic DNA using primers flanking the loxP sites (Fig. 2a) amplified the expected 2490 bp region encompassing the reporter gene (Fig. 2b). Following 2 h of heat shock at 38 °C, a shorter band of 151 bp was observed (Fig. 2a, b). This indicated that the Cre recombinase can excise the reporter gene between the two loxP sites. Of the 14 independent wheat lines tested with this assay, all those which expressed the GUS reporter gene also demonstrated successful construct excision following seedling heat shock.Fig. 2

Heat shock induces excision of the reporter gene. a Excision of the GUS gene and expression of NAM-B1 occurs following heat-shock and expression of Cre recombinase. The location of the primers used to validate the GUS excision are shown in black, alongside the expected band size. b Gel image shows the resultant DNA bands before (blue) and after (yellow) 2 h of heat shock at 38 °C was applied to wheat seedlings. This representative figure depicts the results from individual plants from three independent lines, 2020-20-02, 2020-2-1, and 2020-5-1, at the T₁ generation

As heat shock treatment leads to the successful excision of the reporter gene, we next investigated the expression levels of the gene of interest following recombination. We selected three independent transgenic lines containing the NAM-B1 construct for further testing: 2020-20-02 (one copy), 2020-2-1 (four copies), and 2020-5-1 (ten copies). We also included the null copy control, 2010-54-01. Three or four individual T₁ plants of each line were treated at the two-leaf stage either with 2 h at 38 °C (heat shock; “HS”) or 2 h at 20 °C (non-heat shock; “NHS”), and were sampled for RNA extraction 2 weeks after the treatment. qRT-PCR carried out on the samples found that there was no significant expression of NAM-B1 in the null copy control under any condition (Fig. 3a), as expected. While there was a statistically significant increase in expression in one of the positive transgenic lines before heat shock (“Pre”) compared to the 0 copy control (4 copy line; Student’s T-test, p < 0.05), the spread of gene expression observed was within the biological background levels observed in the 0 copy line. This suggests that no substantial gene expression is observed before heat shock application. However, after heat shock (“Post”) all three lines showed a significant (2020-2-1, 2020-5-1, p < 0.05) or near-significant (2020-20-02, p ~ 0.09) increase in NAM-B1 above the baseline expression before treatment.Fig. 3

Heat shock induces gene expression in a duration-dependent manner. a Heat shock (HS) at 38 °C for 2 h is sufficient to induce expression of NAM-B1 in wheat at the two-leaf stage, while no significant expression is seen in untreated (NHS) plants. Wheat plants were sampled and RNA extracted 2 weeks after heat shock treatment. b No eGFP expression (ii, green) is observed in leaf tissue of untreated (No HS) 1-week old barley seedlings but is visible in all cells following a 30 min heat shock treatment (c). Both the fourth leaf primordia (d) and the second leaf primordia (e) of barley seedlings show mCHERRY expression (i) but no eGFP expression (ii) before heat shock. After 5 min of heat shock, all cells in the fourth leaf primordia (f) show eGFP while sectorial expression of eGFP is observed in the third leaf primordia (g). No expression of eGFP is observed in the second leaf primordia after 5 min of heat shock (h). However, after 15 min of heat shock, eGFP is expressed in all cells of the meristem and second leaf primordia (i). All barley plants were imaged 3 days after treatment. In each pair of images, the reporter gene mCHERRY is shown in panel i (magenta), and eGFP is shown in panel ii (green). The reporter gene mCHERRY (magenta) is present in all samples. Scale bars are 100 µm. The statistical comparison in a was carried out using the Student’s T-test, where N is 3 or 4 for each condition; ns, p > 0.1, ·, p < 0.1, *p < 0.05, **p < 0.001. Gene expression in a was determined using qRT-PCR and is reported relative to the housekeeping gene TaActin (RE_Actin)

We then tested whether the heat shock treatment was able to induce expression of eGFP in the barley transgenics, using progeny of the T₀ lines 00899-04-01 and 00899-01-01. A heat treatment of 30 min of 38 °C induced high levels of eGFP in the leaf tissue of 1-week old seedlings, whereas no background expression of eGFP was observed in the untreated controls (Fig. 3b, c; Additional file 2: Fig. S5A–C). Residual expression of mCHERRY is observed in the leaf tissue following heat shock, though this is substantially reduced compared to the untreated control (Fig. 3b, c; Additional file 2: Fig. S5A–C). This will be due in part to the presence of up to two copies of the construct in the lines, as excision of only one copy is required to see eGFP signal. The mCHERRY protein is also known to be highly stable [27, 28], and thus some residual protein is likely to remain in the 4 days between construct excision and tissue imaging.

Having shown that the heat shock constructs were functional in both barley and wheat, we then investigated how quickly the induction of the reporter genes can occur following heat shock. Using the single copy wheat line, 2020-20-02, and the homozygous two copy barley line, 00899-04-01, we tested the levels of gene expression at 2, 4, 12, and 24 h following heat shock application. We found that in wheat seedlings, 2 h after applying a 2-h heat shock treatment at 38 °C we already saw a significant induction in NAM-B1 expression (p < 0.05, Additional file 2: Fig. S6). There was no further increase in gene expression at the later timepoints, indicating that within 2 h after the completion of the heat shock treatment construct excision had been completed. In barley seedlings the eGFP accumulation at each timepoint following the application of a 30 min heat shock was observed using confocal microscopy (Additional file 2: Fig. S7). We see a gradual increase in eGFP levels from 12 h post-heat shock, continuing until at least 48 h after treatment. No further increase in eGFP levels is observed between 48 h and 8 days after treatment.

Having demonstrated that application of heat shock is sufficient to induce expression of the gene of interest following excision of the reporter gene, we then investigated whether reducing the duration of heat shock treatment could also cause recombination and gene expression. We visualised the expression of eGFP in 1-week-old barley seedlings from line 00899-04-01 grown on filter paper following the application of a 38 °C heat shock of varying durations. To assess the effectiveness of the heat shock treatment at penetrating tissue layers, we analysed the eGFP expression in successive leaves. In grasses each leaf primordium (denoted by “P#”, where “P0” describes the primordium most recently initiated at the meristem) encircles the meristem, and any younger leaves that form. Therefore, each leaf primordium forms a new layer through which heat has to penetrate to reach the innermost tissues, such as young leaf primordia and the meristem itself.

No expression of eGFP was observed without heat shock, as expected (Fig. 3d, e). Five minutes of heat shock was sufficient to induce complete eGFP expression in all outer leaves up to the fourth primordium (P4, Fig. 3f). Single cells expressing eGFP were observed in the third primordium (P3) of plants heat-treated for 5 min (Fig. 3g), while no eGFP was observed in the innermost P2 primordium (Fig. 3h). A 15-min heat treatment was sufficient to induce eGFP expression in all leaf layers and tissues, including the second leaf primordium (P2) and the meristem (Fig. 3i), in 1-week-old seedlings. We validated this result in the second barley transgenic line 00899-01-01, confirming that a 15-min heat treatment would be sufficient to induce eGFP expression in both leaf and primordia tissues (Additional file 2: Figure S5D, E). This illustrates the tunability of expression for different tissue layers based on heat shock duration.

After establishing that the applied heat shock was sufficient to induce gene expression in treated seedlings, we next determined the extent to which this was true at later developmental stages. We found that a 2-h application of heat shock at 38 °C at the three-leaf stage in soil-grown barley was sufficient to induce gene expression in the outer leaves (i.e. leaf 3, Fig. 4a, b). However, minimal eGFP expression was observed in later developing leaves (i.e. leaf 5, Fig. 4c, d) and no eGFP was observed in the inflorescence meristem (Fig. 4e, f). This suggests that the heat treatment was not able to fully penetrate to the meristematic cells at the three-leaf stage of soil grown plants, resulting in new tissue growth containing only the unexcised construct.Fig. 4

Testing heat shock stages. Heat shock treatments of plants at different developmental stages: 3 leaf stage (a–f), flag leaf emergence (g–l), and ear emergence (m–r). Heat shock was carried out at 38 °C for 2 h, and each panel shows the mCHERRY expression (i, magenta) and eGFP expression (ii, green). Tissues are imaged in pairs, showing the equivalent tissue without (left) and with (right) heat shock. a, b Leaf 3, c, d Leaf 5, e, f inflorescence meristem. g, h Flag leaf blade, i, j developing florets, k, l young awns. m, n stem epidermis below the inflorescence, o, p mature awn, q, r stigma. Scale bars 100 µm. At all developmental stages, plants were imaged four days after treatment

We then investigated whether this result would be recapitulated in the wheat transgenics. Soil-grown plants from lines 2020-20-02, 2020-2-1, and 2020-19-01 (n = 5 for each line) were treated with a 2 h 38 °C heat shock at the three-leaf stage. Similarly to barley, we found that tissue from leaf 3 showed complete excision of the reporter gene (Additional file 2: Figure S8A, “3rd”), while tissue from leaf 5 showed inconsistent excision between independent replicates (Additional file 2: Figure S8A, “5th”). Some individuals (6/15) showed full excision in the leaf 5 tissue (Additional file 2: Figure S8A, left panel), indicating that the meristematic cells from which leaf 5 developed had experienced sufficient heat shock for full excision of the reporter gene. Others showed an intermediate state (4/15), with the presence of both the full construct, at 2490 bp, and the excised construct, at 151 bp (Additional file 2: Figure S8A, middle panel). In these cases, it is likely that the fifth leaf developed from a mixture of excised and unexcised primordia cells, leading to a mosaic in the mature tissue. Finally, some individuals showed no evidence of construct excision (5/15), indicating that the fifth leaf primordia cells were not successfully heat shocked (Additional file 2: Figure S8A, right panel).

As a 2-h heat shock treatment at 38 °C was not sufficient to consistently induce construct excision in the entirety of the three-leaf stage plant, we decided to investigate whether this treatment would be able to induce construct excision at later developmental stages. At the flag leaf emergence stage, we found that this treatment was sufficient to induce complete construct excision in all aerial tissues in barley (Fig. 4g–l). eGFP expression was induced in the flag leaf (Fig. 4g, h), the developing florets (Fig. 4i, j), and in the young awns (Fig. 4k, l). This was mirrored in the wheat transgenics, where the same heat shock treatment to lines 2020-2-1 and 2020-19-01 (n = 6 for each line) leads to activation of the construct and excision of the reporter gene in both the flag leaf and spikelet tissue (Additional file 2: Figure S8B). In two of the 12 samples, the excision in the basal spikelet was not complete, though the excised construct is also present (Additional file 2: Figure S8B; purple asterisk). We also found that the heat shock treatment of the barley plants at spike emergence was also sufficient to induce eGFP expression in all aerial organs, including in the stem and the floral organs of the plant (Fig. 4m–r).

As treatment at 38 °C can induce expression of the heat-shock driven construct, we then investigated whether temperatures lower than 38 °C could lead to inadvertent (i.e. “leaky”) activation of the construct. We focussed on the barley transgenic system, as the use of fluorescent markers allowed the visualisation of cell-specific changes in expression. In 1-week old seedlings, treated with a heat shock of 30 min, we only observed high levels of eGFP expression at 38 °C or higher (Fig. 5m–t). This was consistent across all three emerged leaves and the meristem. At 32 °C, some individual cells exhibited eGFP expression in the emerged leaves (Fig. 5i–k), though no eGFP was observed in the meristem or leaf primordia (Fig. 5l). Even fewer cells expressed eGFP at 26 °C, and only in leaves 1 and 2 (Fig. 5e–h). No expression of eGFP was observed at 20 °C, in keeping with previous experiments (Fig. 5a–d). This was also consistent with previous studies of the HvHSP17 promoter [10].Fig. 5

eGFP is strongly induced only at or above 38 °C. a–t One-week old barley seedings were subjected to a heat shock of 30 min, at one of five different temperatures: 20 °C (a–d), 26 °C (e–h), 32 °C (i–l), 38 °C (m–p), and 44 °C (q–t). The mCHERRY (i, magenta) and eGFP (ii, green) expression is shown for each tissue sample. Images of leaf 1 (a, e, i, m, q), leaf 2 (b, f, j, n, r), leaf 3 (c, g, k, o, s), and the meristem with attached leaf primordia (d, h, l, p, t) are shown for each temperature treatment. All plants were imaged 3 days following treatment. The scale bar is 100 µm. In 1-week old wheat seedlings, NAM-B1 is induced most strongly at 38 °C, though also to a lesser extent at 32 °C in the single copy line 2020-20-02 (u). Statistical comparisons were carried out using the Wilcox test due to the non-normality of the data, with N = 4 for all conditions; *p < 0.05, ns, p > 0.05. Gene expression in u was determined using qRT-PCR and is reported relative to the housekeeping gene TaActin (RE_Actin)

We then tested whether the induction of the heat-shock construct at 32 °C was visible in the wheat transgenic lines, using 1-week old T₁ plants descended from the single-copy line 2020-20-02. Ten days after a 2-h heat shock, qRT-PCR of seedling leaf tissue showed a significant increase in NAM-B1 expression at 38 °C, as expected (Fig. 5u, p < 0.05). Those treated with a 2-h 32 °C heat shock also showed a significant increase in NAM-B1 expression (Fig. 5u, p < 0.05). However, the level of expression at 32 °C was significantly lower than that observed after heat shock at 38 °C (p < 0.05). This mirrors the comparatively low levels of eGFP expression observed at 32 °C in the barley transgenics.

To prevent expression of the Cre recombinase during cloning in E. coli, we introduced the Arabidopsis thaliana U5 small nuclear ribonucleoprotein component intron (from pICSL80006, TSL SynBio) at 254 bp in the coding sequence, based on published mammalian studies [42]. The Cre recombinase gene was amplified in two sections, using primer sets designed using the general format: NGAAGACNN + BpiI-specific 4 bp overhang + 18–30 bp of the target sequence ([40]; Additional file 2: Table S3), which introduced flanking BpiI sites and allowed ligation into the pICH41308 backbone. Simultaneously, the U5 intron sequence was amplified to allow ligation between the two Cre recombinase sequence fragments. 2 µL of each PCR fragment was included in the ligation reaction, following the protocol described in [38]. The resulting sequence is provided in Additional file 1.

The domesticated NAM-B1 sequence, and the sequence of the ZmUbi promoter, along with its 5′UTR (from pICSL12009 [43]) were cloned directly into the universal acceptor pUAP1 [40] to generate level 0 constructs with the desired 4 base pair overhangs, hereafter pNAM-B1 and pZmUbi, respectively (see primers used in Additional file 2: Table S3). The PCR-amplified sequences were ligated into pUAP1 using the BpiI enzyme, following the protocol from [38].

To introduce flanking loxP sites around the reporter gene and corresponding terminator, an additional step was required alongside the standard Golden Gate cloning protocol. The short BsaI protocol was carried out as described by TSL Synbio, using the loxP level 0.5 construct developed by ENSA (EC10161) (Fig. 6A). For the wheat construct, the GUS coding sequence from TSL Synbio (pICH75111) and the nosT terminator sequence (pICH41421) were cloned into the loxP construct (Additional file 2: Figure S2A). For the barley construct, an ER targeting sequence (EC71090), the mCHERRY coding sequence (EC71091), the HDEL sequence (EC71020), and the 35S terminator (EC41414) were cloned into the loxP construct (Additional file 2: Figure S3A). The ER targeting sequence and the HDEL sequence used are provided in Additional file 1. This created a “level 0.5” vector, hereafter referred to as ploxP-GUS or ploxP-mCHERRY, with loxP sites flanking the coding region and terminator. Unlike standard level 0 constructs, the level 0.5 constructs contain flanking Esp3I sites, rather than BsaI sites. The level 0.5 construct combines positions B1 and B2 of the standard Type IIS cloning syntax, allowing its insertion upstream of standard coding regions, and downstream of standard promoter regions [40].Fig. 6

Assembly of the Level 0.5 and Level 1 vectors. a The reporter gene (Gene 1) and its associated terminator (“T1”) were cloned into the Level 0 vector EC10161, using the Type IIS restriction enzyme BsaI. The resulting Level 0.5 vector contains loxP sites flanking the inserted reporter gene and terminator. b The desired promoter sequence, loxP-flanked reporter gene, gene of interest (Gene 2), and corresponding terminator (T2) are cloned into a Level 1, position 3 vector (e.g. pAGM8031) in a reaction involving both BsaI, for the canonical Level 0 and Level 1 parts, and Esp3I, for the loxP-containing Level 0.5 vector (see a). In both cases, additional components can be included during the cloning step, such as targeting sequences

For each system, a level 1 construct was synthesized that combined a constitutive promoter, the loxP-flanked reporter gene, and the gene of interest (Fig. 6b). For the wheat construct, the promoter construct (pZmUbi), the loxP construct (ploxP-GUS), the gene of interest (pNAM-B1), and the nosT terminator (pICH41421) were then cloned into the level 1 backbone, position 3 (pICH47751) (Additional file 2: Figure S2B). For the barley construct, the promoter construct (EC71139), the loxP construct (ploxP-mCHERRY), the ER targeting sequence (EC71090), the gene of interest (EC71088), the HDEL sequence (EC71020), and the actin terminator (EC44300) were then cloned into a level 1 backbone, position 3 (EC47822) (Additional file 2: Figure S3B). In both cases, this step used a modified version of the BsaI protocol from [38] which contains 10 U BsaI and 10 U Esp3I, instead of 20 U of BsaI, to allow insertion of the loxP construct (Fig. 1b).

The level 1 constructs were cloned into the appropriate final backbone (level 2/M), integrating the loxP construct with the Cre recombinase construct, and the Hygromycin selection cassette, flanked by right and left border repeats for Agrobacterium-mediated transformation. For the wheat construct, the NAM-B1-loxP level 1 construct was cloned into the level M backbone (pAGM8031), alongside a selection cassette (pOsAct::Hygromycin, pICH47802), the pHvHsp17::Cre construct (EC71173), and the appropriate end linker (pICH50892, Additional file 2: Figure S2C). For the barley construct, the level 1 eGFP construct, the pHvHSP17::Cre construct (EC71173), and the appropriate end linker (EC41766) were cloned into the level 2 plasmid EC15027, which contains an appropriate Hygromycin expression system (Additional file 2: Figure S3C). Cloning was carried out using the BbsI protocol from [38]. These constructs will be referred to as HS_NAM-B1 and HS_GFP, respectively. The sequences of all constructs at all stages were verified using Sanger sequencing (Eurofins). The specific cloning steps for the wheat and barley constructs, respectively, are illustrated in Additional file 2: Figures S2 and S3.

Seeds were germinated at 4 °C on damp filter paper for 48 h in the dark before they were moved to room temperature and were grown under ambient light conditions. The seedlings were grown for 5 to 7 days following germination before the heat shock treatment. Half of the seedlings were treated to a two-hour heat shock at 38 °C in a temperature-controlled Ecotron incubator (Infors HT). The remaining seedlings were left at ambient room temperature during this period.

Seeds were germinated at 4 °C in the dark for 48 h on damp filter paper, and then allowed to grow for a further 48 h under ambient light conditions. The germinated seedlings were then sown into P96 trays containing 85% fine peat with 15% horticultural grit. They were grown until the 2–3 leaf stage before transplanting into 1 L pots. The plants experienced cycles of 16 h of light at 20 °C and 8 h of dark at 15 °C at 60% relative humidity. Plants received approximately 450–550 µmols of photons at flowering height.

Where required, a heat shock treatment was applied to plants grown in soil. Depending on the developmental stage of the plant, either a P96 tray (3-leaf stage or earlier) or up to nine 1 L pots (flag leaf emergence or later) were taken to a controlled environment chamber (Sanyo MLR-351H) and incubated at 38 °C for 2 h, at 85% relative humidity and light level of 270 µmols of photons, before being returned to the original environment. A second set of non-heat shock controls were left in the original environment for the period of the heat shock.

T₀ plants were grown in 1 L pots in John Innes cereal mix in controlled environment rooms (Gallen Kamp) at 75% humidity, with a day/night cycle of 16 h light at 15 °C, 8 h dark at 12 °C. T₁–T₃ generations were grown in 1 L pots of cereal mix in temperature-controlled greenhouse conditions, at 20 °C during the day and 15 °C during the night, and allowed to self to bulk seed.

Leaf tissue was sampled from individual plants and snap frozen in liquid N₂. The stage at which the tissue was sampled depended on the experiment in question. Initial validation of construct expression was carried out on seedling leaf samples, taken between 5- and 7-days following germination from the first true leaf of the plant (i.e. second leaf stage). Tissue samples from plants grown on soil in controlled environment conditions were taken at the third leaf stage. In all cases, approximately 3 cm of tissue were sampled from the tip of the leaf. The snap-frozen tissue was then ground to a fine powder and RNAs were extracted using TRIzol^® Reagent (ThermoFisher). cDNA synthesis was carried out using the Invitrogen M-MLV reverse transcriptase, following treatment with RQ1 RNase-Free DNase (Promega).

Initial validation of Cre recombinase expression was carried out using RT-PCR on cDNA samples from lines carrying the heat-shock construct (HS_NAM-B1). Touchdown PCR was carried out on the relevant samples using NEB Taq polymerase using the primer pair SH002/SH053 (Additional file 2: Table S3) under the following conditions: 5 min at 95 °C, 10 cycles of 30 s at 95 °C, 1 min starting at 62 °C and decreasing by 1° per cycle, and 1 min at 68 °C, a further 25 cycles of 30 s at 95 °C, 1 min at 52 °C, and 1 min at 68 °C, followed by a final extension at 68 °C.

The expression of NAM-B1 in the wheat transgenic lines was validated using quantitative RT-PCR. Primers specific to the domesticated version of NAM-B1 present in the construct were designed and used (SH049/SH051) alongside previously published primers for the internal control gene TaActin ([5]; Additional file 2: Table S3). Efficiencies for the primers were calculated using pooled cDNA from samples which were known to be expressing the transgenic construct (Additional file 2: Table S3). qRT-PCR reactions were performed using the LightCycler^® 480 SYBR Green I Master Mix with a LightCycler 480 instrument (Roche Applied Science, UK) under the following conditions: 5 min at 95 °C; 45 cycles of 10 s at 95 °C, 15 s at 60 °C, 30 s at 72 °C; followed by a dissociation curve from 60 to 95 °C to determine primer specificity. In all cases, three technical replicates were carried out per sample and the expression of NAM-B1 was recorded relative to TaActin.