Background

Hop (Humulus lupulus L.) is a diploid, dioecious, perennial climbing plant belonging to the Cannabaceae family. Female plants are cultivated for the commercial production of inflorescences (cones), which are mainly used in the brewing industry but also commonly used in the production of pharmaceuticals and cosmetics [1]. In cones, flavour-active secondary metabolites are biosynthesized and accumulate in glandular trichomes and lupulin glands. There are three key classes of natural products: alpha and beta bitter acids (humulone and lupulone derivatives, respectively), prenylated flavonoids (primarily xanthohumol and desmethylxanthohumol) and essential oils composed mainly of myrcene, α-humulene and β-caryophyllene [2].

The contents of alpha and beta bitter acids in hop cones (5 to 30% of the dry weight) are major chemical characteristics and economical traits of different cultivars during hop production. Their biosynthesis requires precursor sources that directly originate from primary sucrose metabolism and amino acid precursors such as leucine, isoleucine, valine and phenylalanine [3]. Phenylalanine is biosynthesized by the shikimate pathway and others by the branched-chain amino acid (BCAA) pathway in chloroplasts. The final step in the BCAA pathway is catalysed by the enzyme branched-chain amino transferase 2 (HlBCAT2). For bitter acid biosynthesis, BCAAs are then degraded in the mitochondria by branched-chain amino transferase 1 (HlBCAT1) and branched-chain keto-acid dehydrogenase (HlBCKDH) to isovaleric, 2-methylbutyric and isobutyric acids, respectively [4]. Coenzyme A is ligated to these precursors by two branched short-chain fatty acid CoA ligases (HlCCL2 and HlCCL4) in the cytosol [5]. Valerophenone synthase (HlVPS) synthesizes phloroisovalerophenone (PIVP), phloro-2-methylbutyrophenone (PMBP) and phloroisobutyrophenone (PIBP) from precursors and three malonyl-CoAs from phenylalanine [6].

Isoprenoids are a large and highly diverse group synthesized from common prenyl diphosphate precursors. Plants synthesize isopentenyl diphosphate (IPP) via mevalonate and methylerythritol phosphate (MEP) pathways [7]. The MEP pathway is more efficient in hop lupulin glands according to transcriptome [4] and proteome [3] studies. Terpenic compounds are the main components of hop essential oils.

Bitter acids and prenylflavonoids are additional specific compounds of prenylation in lupulin glands. Both are prenylated by a functional heterocomplex of two aromatic prenyltransferases, HlPT1L and HlPT2, in chloroplasts [8,9,10]. HlPT1L catalyses the first prenylation step, and HlPT2 catalyses the next two prenylation steps. The first step produces alpha bitter acid precursors, such as deoxyhumulone, deoxyadhumulone and deoxycohumulone, and the second step produces beta bitter acids, such as lupulone, adlupulone and colupulone. Alpha bitter acids (humulone, adhumulone and cohumulone) are finally produced by humulone synthase (HlHS1 and HlHS2) via the oxygenation of deoxy-precursors [3]. The content of individual analogues of alpha and beta bitter acids is determined by the ratios of precursors and is genetically imprinted in the genomes of individual genotypes; for example, cohumulone (from 12 to 55% of alpha acids) and colupulone (from 30 to 80% of beta acids) contents are very stable and heritable traits [2].

Since the expression of biosynthetic genes is the highest in lupulin glands, there is a tissue-specific regulatory network of transcription factors mainly from the bHLH, MYB, WDR and WRKY families [1]. Promoters of bitter acid and prenylflavonoid biosynthesis pathway genes are regulated by heterotrimeric ternary MBW (HlMyb3/HlbHLH2/HlWDR1 or HlMYB2/HlbHLH2/HlWDR1) complexes [11], binary (HlbHLH2/HlWDR1 or HlWRKY1/HlWDR1) transcription activation complexes [11, 12] or individual transcription factors (HlWRKY1 or HlMyb8) [13, 14]. This regulation allows hop plants to react to developmental changes, environmental stresses and weather conditions, which influence bitter acid contents.

The bitter acid content in hop cones varies across hop cultivars due to their genetic background. Hop cultivars are divided into aromatic hops, with low alpha bitter acid contents in dry cones ranging from 0.5 to 8%, and bitter hops, with alpha bitter acid contents in dry cones over 9 to 23% [15]. It was found that the content of alpha bitter acids very strongly and significantly correlated with the number and size of lupulin glands in hop cones [2, 16, 17]. Therefore, bitter acid contents can be increased either by increasing their production inside secretory cells or by increasing the density of glands on individual cones. The development of multicellular glandular trichomes includes the enlargement of single epidermal cells, followed by several polarized and localized cell divisions and remodelling to generate branched or unbranched structures perpendicular to the epidermal surface. The lupulin glands develop on the abaxial surface of the leaf, bract or bracteole primordia and continue to form until expansion stops [1]. There are already molecular data for genes that play a specific role in glandular trichome development, especially transcription factors, cell cycle regulators, and receptors involved in phytohormone-induced signalling cascades [18].

The model plant A. thaliana is the most utilized for the developmental study of trichomes, which are single-celled and non-glandular [19]. The current model includes transcription factors that function as positive or negative regulators of trichome formation, upstream regulators of these two groups of regulators, and downstream components [20]. The positive regulators belong to three protein classes and include a WD40-repeat protein TRANSPARENT TESTA GLABRA1 (TTG1), three R2R3 MYB-related transcription factors GLABRA1 (GL1, MYB23, MYB5) and four basic helix-loop-helix (bHLH)-like transcription factors GLABRA3 (GL3), ENHANCER OF GLABRA3 (EGL3), TRANSPARENT TESTA (TT8), and MYC-1. They act partially redundantly and form a multimeric activator complex, also known as the MYB-bHLH-WD40 (MBW) complex, which binds the promoter of GLABRA2 (GL2) and several homeodomain (HD-ZIP class IV) transcription factors [19]. GL2 encodes a homeodomain protein required for subsequent phases of trichome morphogenesis, such as endoreduplication, branching, and maturation of the cell wall. The MBW activator complex is usually negatively regulated by single-repeat R3 MYBs, such as CAPRICE (CPC), TRIPTYCHON (TRY), ENHANCER OF TRY and CPC1 (ETC1), ETC2, TRICHOMELESS1 (TCL1) and TCL2/CPL4 [20]. Additionally, several C2H2 zinc finger protein transcription factor genes, including GLABROUS INFLORESCENCE STEMS (GIS, GIS2) and ZINC FINGER PROTEIN 8 and 5 (ZFP8, ZFP5), controlling GL1 and GL3 of the core MBW complex, have also been identified [19, 20].

In contrast, knowledge of the development of glandular trichomes is very limited, and some data have shows that glandular trichome pathways are not as conserved as the nonglandular trichome pathways known in Arabidopsis. What we know about glandular trichome development is mostly from work on Artemisia, tomato and tobacco [19].

In tomato, several genes have been reported that are required for the proper development and function of different types of glandular trichomes. The initiation and development of type I trichomes was controlled by a woolly (Wo) gene encoding a class IV homeodomain-leucine zipper protein homologue of Arabidopsis GL2 and a B-type cyclin gene, SlCycB2 (possibly regulated by Wo). A hairy phenotype was caused by the overproduction of mutant alleles of Wo in type I trichomes. However, the suppression of Wo or SlCycB2 expression by RNA interference decreased trichome density in tomato [21, 21, 22], and its homologue HlCYCB2–4 was also upregulated in young cones of bitter hop cultivars.

The previously identified lupulin-upregulated homeobox-leucine zipper protein HlHB51 was also upregulated in young cones of bitter hop cultivars [1]. ATHB-51 interacts with the meristem regulator LEAFY and participates in trichome formation, leaf morphogenesis and floral meristem determinacy in Arabidopsis [30], and it is an essential regulator of multicellular trichome development in Cucumis sativum [31]. Similarly expressed homeobox-leucine zipper proteins HlZHD6, HlHB14 and HlCYCNB1 probably participate in meristem determinacy and development. HlHDG11 was similarly upregulated in lupulin glands in a previous study [1]. This gene is involved in epidermal cell differentiation, and HDG11 mutants show excess branching of the trichome in A. thaliana [32].

In a previous study, the axial regulator HlYABBY1 was significantly downregulated in lupulin glands and young cones [1], as in our results. We can suppose from its upregulation in young cones of bitter hop cultivars that this gene may be involved in lupulin gland development at abaxial sites of bracts and bracteoles [33].

The MIXTA transcription factor is also considered a positive regulator of glandular trichome formation [19, 20]. It was obvious from our results for its homologue HlMYB106 that this gene is a general transcription factor for glandular trichome development.

HlWRKY44, TRANSPARENT TESTA GLABRA2 (TTG2) factor, regulates trichome-specific TTG1, GL2, GL3 and TRY factors [34]. Its downregulation in the lupulin glands of bitter hop cultivars probably results in downregulation of the MYB-related transcription factor TRY, a negative regulator of trichome development. However, WRKY44 is also involved in the biosynthesis of proanthocyanins in fruits [35] or mucilage and tannins in seed coats [36].

The homologue of the bHLH factor HlGL3 was downregulated in lupulin glands without differences between aromatic and bitter hop cultivars. Therefore, we cannot confirm that this gene is a part of the MYB-bHLH-WD40 complex, and thorough combinatorial analysis will be necessary [11].

Expression differences in hop cone tissues were previously found [1] for AP2/ERF and the B3 domain-containing transcription factor HlRAV1, which were confirmed despite a lack of influence on glandular trichome density or alpha bitter acid content. Other genes that were overexpressed in lupulin glands include the Abl interactor-like proteins HlABIL2 or HlABIL3 [1], which are involved in the regulation of actin and microtubule organization as part of a SCAR/WAVE complex that activates the ARP2/3 complex [37]. Based on the expression differences in apexes, flowers, young cones or lupulin glands, we cannot exactly determine what aspect of trichome, cone or seed development is influenced, even if these genes play a role in the development of plant cell shape and can cause a distortion in the trichome phenotype [38].

Conclusions

Gene expression analyses enabled us to identify differences between aromatic and bitter hops. This study confirmed that the bitter acid content in glandular trichomes (lupulin glands) is dependent on the last step of alpha bitter acid biosynthesis and glandular trichome density. Humulone synthase genes (two analogues) are important for alpha bitter acid content in the glandular trichomes of hop cones. Differential gene expression analyses showed that the MYB transcription factors HlETC1, HlMYB61 and HlMYB5, homeobox-leucine zipper protein HlGLABRA 2, C2H2 zinc finger protein HlZFP8 and axial regulator HlYABBY1 may be key regulatory factors for lupulin gland density and morphological and developmental differentiation of glandular trichomes inside hop cones. Analyses also showed that glandular trichome initiation and development are controlled by a large regulatory network and should be studied in detail in the future.

Methods

Plant materials

The hop plants used were grown under standard agronomic conditions in experimental fields on the Steknik farm of the Hop Research Institute in Zatec (Saaz), CR. Based on general long-term knowledge, we selected the “aromatic” hop cultivars Saaz, Fuggle, Hallertauer and Kazbek (alpha bitter acid contents varying from 2.5 to 8.0% and beta bitter acid contents varying from 2.0 to 6.0% in cones) and the “bitter” hop cultivars Vital, Herkules, Columbus, and Magnum (alpha bitter acid contents varying from 11.0 to 18.0% and beta bitter acid contents varying from 4.0 to 10.0% in cones) as contrasting hop genotypes for group analyses [2]. In 2020, three pooled samples of eight plants for each hop cultivar (Saaz, Fuggle, Hallertauer, Vital, Herkules, and Columbus) were collected from different tissues during development: apexes in April (only Saaz and Vital), flowers in July, leaves and young cones in August, and mature cones in August and September. Samples collected in August 2016 were used to increase the tissue variability. Three pooled samples containing one kilogram fresh weight of mature cones were used for bract and lupulin gland analyses in the hop cultivars Saaz, Fuggle, Hallertauer, Kazbek, Vital, Columbus and Magnum. Three pooled samples of four plants of the hop cultivars Saaz, Fuggle Columbus and Magnum collected in May 2013 were used to increase the variability of leaf and apex tissues. All samples were immediately frozen in liquid nitrogen and stored in a deep freezer (− 80 °C). Lupulin glands were separated from mature cones by agitation in liquid nitrogen followed by filtration through a 1 mm metal sieve to remove cone debris [1]. The glands were recovered from liquid nitrogen, and the rest of the cones were analysed as bracts without lupulin. We did not measure either the bitter acid contents or numbers of lupulin glands in hop cone samples.

RNA isolation and gene expression analyses

RNA was isolated using PureLink™ Plant RNA Reagent (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol and purified by DNaseI treatment on a column (RNeasy Plant Mini Kit, Qiagen, Hilden, FRG) [11]. RNA samples were reverse transcribed by oligo (dT)18 primer and a First Strand cDNA Synthesis Kit (Roche Diagnostics, Mannheim, Germany) at 50 °C for 60 min. NGS genome information [39] and the HopBase database [38, 40] were used to search for bitter acid biosynthesis [3] and trichome-specific transcription factor [1] gene sequences (Table 1). Genes were selected based on previous works [1, 3,4,5,6,7,8,9,10,11,12,13,14, 18,19,20]. Advanced BLAST 2.0 (http://www.ncbi.nlm.nih.gov/blast/blast.cgi) and the MegAlign module (LASERGENE system v. 7.1, DNAStar, Madison, WI, USA) were used to evaluate sequence homology. RealTimeDesign software (LGC Biosearch Technologies, Petaluma, CA, USA) was used to design real-time PCR primers (Table S1). A total of 2 μl of 50 × diluted cDNA was used in a 20 μl PCR with iTaq universal SYBR green Supermix (Bio–Rad Laboratories, Hercules, CA, USA) in a CFX Connect real-time PCR cycler (Bio–Rad Laboratories, Hercules, CA, USA). Five reference genes (Table 1), which were found to be constitutively expressed in different tissues [11] by NormFinder in GenEx v.6.0.1.612 (MultiD Analyzes AB, Gothenburg, Sweden), were used for normalization of the samples. The relative expression of five reference genes was calculated by the “delta-delta method” (RE = 2-∆CT) for each sample. SigmaPlot for Windows v.10.0.0.54 (Systat Software Inc., San Jose, CA) was used for statistical group (average, standard deviation) and unpaired two-sample t-test analyses of relative expression.

Table 1 List of analysed bitter acid biosynthesis gene, trichome-specific transcription factor and reference gene sequences
Full size table