Dev Genes Evol (2010) 220:25–40 DOI 10.1007/s00427-010-0326-4

ORIGINAL ARTICLE

A complex case of simple leaves: indeterminate leaves co-express ARP and KNOX1 genes

Kanae Nishii & Michael Möller & Catherine Kidner & Alberto Spada & Raffaella Mantegazza & Chun-Neng Wang & Toshiyuki Nagata

Received: 27 November 2009 /Accepted: 15 April 2010 /Published online: 26 May 2010 # Springer-Verlag 2010

Abstract The mutually exclusive relationship between Here, we present several datasets illustrating the co- ARP and KNOX1 genes in the shoot apical meristem and expression of ARP and KNOX1 genes in the shoot apical leaf primordia in simple leaved such as Arabidopsis meristem, leaf primordia, and developing leaves in plants has been well characterized. Overlapping expression with simple leaves and simple primordia. domains of these genes in leaf primordia have been plants produce unequal cotyledons due to the continued described for many compound leaved plants such as activity of a basal meristem and produce foliar leaves Solanum lycopersicum and Cardamine hirsuta and are termed “phyllomorphs” from the groove meristem in the regarded as a characteristic of compound leaved plants. acaulescent species Streptocarpus rexii and leaves from a shoot apical meristem in the caulescent Streptocarpus Communicated by K. Schneitz glandulosissimus. We demonstrate that the simple leaves Electronic supplementary material The online version of this article in both species possess a greatly extended basal meriste- (doi:10.1007/s00427-010-0326-4) contains supplementary material, matic activity that persists over most of the leaf’s growth. which is available to authorized users. The area of basal meristem activity coincides with the co- K. Nishii (*) : C.-N. Wang expression domain of ARP and KNOX1 genes. We suggest Institute of Ecology and Evolutionary Biology, that the co-expression of ARP and KNOX1 genes is not Department of Life Science, National Taiwan University, exclusive to compound leaved plants but is associated with Rm. 1207, Life Science Building, No.1, Sec. 4, Roosevelt Road, foci of meristematic activity in leaves. Taipei 10617 Taiwan, Republic of China e-mail: [email protected] . . . . : Keywords ARP KNOX1 Streptocarpus M. Möller (*) C. Kidner Meristem Royal Botanic Garden Edinburgh, 20A Inverleith Row, Edinburgh EH3 5LR, UK e-mail: [email protected] Introduction : A. Spada R. Mantegazza In most seed plants, the above ground parts of plants are Department of Biology, University of Milan, Via Celoria 26, formed from layered shoot apical meristems (SAMs). Milan 20133, Italy SAMs are composed of three domains that fulfill distinct functions: the central zone contains a self-renewing pool of C. Kidner stem cells, the peripheral zone from which cells are Institute of Molecular Sciences, University of Edinburgh, King’s Buildings, Mayfield Road, recruited into developing organs, and the rib zone, which Edinburgh EH9 3JR, UK is the origin of ground tissue within the stem (Gifford and Corson 1971; Bowman and Eshed 2000;Carlesand T. Nagata Fletcher 2003). Leaf primordia arise from the peripheral Faculty of Bioscience and Applied Chemistry, Hosei University, 3-7-2 Kajino-cho, Koganei-shi, zone of the SAM. During leaf development, cell division Tokyo 184-8584, Japan initially occurs throughout the primordia, but soon after- 26 Dev Genes Evol (2010) 220:25–40 wards becomes limited to the base of the leaf and finally (Jong 1970; Jong and Burtt 1975). Three meristems, the cell expansion alone is responsible for the continuing “basal meristem”, “petiolode meristem”,and“groove enlargement of the leaf (Poethig and Sussex 1985a, meristem” are involved in the production of a leaf-like 1985b; Nath et al. 2003). organ termed “phyllomorph” (Fig. 1). The basal meristem The underlying genetic pathways for SAM establish- is found at the base of the lamina and is a region of ment, maintenance, and leaf development have been persistent cell division supplying new laminar tissue. The extensively studied in a number of model plants such as petiolode meristem, a diffuse rib meristematic area of the Arabidopsis, Antirrhinum, and maize. In Arabidopsis, “petiolode” (stem-like petiole of the leaf), is responsible for WUSCHEL (WUS) and CLAVATA (CLV) are key genes that petiolode and midrib extension. The groove meristem is act in the central zone, specifying the fate of the stem cells positioned on the petiolode and responsible for the and maintaining a relatively constant cell number in the formation of new organs, phyllomorphs or inflorescences SAM (Mayer et al. 1998; Fletcher et al. 1999; Schoof et al. (Fig. 1) (see also Fig. 1 in Mantegazza et al. 2007; Jong 2000). Class 1 KNOTTED-like homeobox genes (KNOX1) 1970, 1978). are also important for maintaining meristem cells in an All Streptocarpus species exhibit anisocotyly, where one undifferentiated state (Vollbrecht et al. 1991; Barton and of the two initially equal-sized cotyledons develops into a Poethig 1993; Long et al. 1996). KNOX1 genes have macrocotyledon through the extended activity of its basal undergone duplication in several different lineages, and meristem shortly after germination to form the first photosyn- studies on expression patterns and mutant phenotypes thetic organ. In acaulescent species (unifoliates and rosulates), suggest that sub- and neo-functionalizations have occurred an organized SAM is not produced between the cotyledons. (Reiser et al. 2000). In , genes of the SHOOTMER- The macrocotyledon becomes a “cotyledonary phyllomorph”, ISTEMLESS (STM)-like KNOX1 clade are required for and rosulate species initiate additional phyllomorphs from the meristem function and are expressed throughout the groove meristem while unifoliate species maintain only a meristem. Genes in the BREVIPEDICELLUS (BP)-like single enlarged cotyledon (Fig. 1) (see also Fig. 1 in Harrison clade are required for meristem function in the absence of et al. 2005a;Jong1970). The basal meristem is active until ASYMMETRIC LEAVES 1 (AS1)andSTM and are the leaves produce an inflorescence at its base, and this can expressed predominantly in the peripheral zone in Arabi- take more than 4 years in some species during which the dopsis (Byrne et al. 2000; Reiser et al. 2000). lamina continues to grow (Hilliard and Burtt 1971). ARP genes, named after the orthologous genes AS1 from Caulescent Streptocarpus species, while also showing Arabidopsis thaliana, ROUGH SHEATH2 (RS2) from anisocotyly and no embryonic SAM, quickly develop a maize and PHANTASTICA (PHAN) from Antirrhinum conventional shoot structure with stem and decussate pairs of majus, are MYB-like transcription factors involved in simple leaves from a central layered SAM produced post- adaxial–abaxial and proximal–distal axis formation during embryogenically (Jong 1970; Imaichi et al. 2007). Leaves of leaf morphogenesis. ARP genes negatively regulate KNOX1 the caulescent Streptocarpus pallidiflorus show a group of genes in determined organ primordia (Waites et al. 1998; small cells at the base, which were compared to a possible Timmermans et al. 1999; Tsiantis et al. 1999; Byrne et al. basal meristem (Imaichi et al. 2007). This may indicate that 2000). In Arabidopsis, STM represses the expression of AS1 the macrocotyledon and foliage leaf development in Strepto- in the SAM, and AS1 in turn represses BP/KNAT1 in leaf carpus share certain features. primordia (Byrne et al. 2002). This antagonistic relationship The activity of several genes, involved in meristem is characteristic of many plants with simple leaves (Waites function in model species, has previously been studied in et al. 1998; Byrne et al. 2002), but has been shown to break Streptocarpus. Immunolocalization of KNOX proteins and down in plants with compound leaves. In these, ARP and KNOX1 SSTM1 (Streptocarpus STM1) RT-PCR showed KNOX1 genes are either co-expressed in leaf primordia KNOX1 expression in the SAM and proximal region of (Cardamine hirsuta; Hay and Tsiantis 2006) or in the SAM leaves, but not in incipient leaf primordia in the SAM in the and leaf primordia (Solanum lycopersicum; Hareven et al. caulescent Streptocarpus saxorum. KNOX proteins were 1996; Chen et al. 1997; Koltai and Bird 2000). detected in inflorescences, groove meristems, and leaf Most species of the family Gesneriaceae have simple primordia of the acaulescent Streptocarpus rexii (Harrison leaves with a simple leaf primordium produced from a et al. 2005a). Further detailed studies on S. rexii revealed SAM (Lai 2001; Barth et al. 2009). Caulescent species in that SrSTM1 (S. rexii STM1) is expressed not only in the the genus Streptocarpus conform to this development. groove meristem, but also in the basal meristem of Acaulescent Streptocarpus species, though simple leaved, cotyledons (Mantegazza et al. 2007). Expression was also show a radically different organization and development of found throughout the developing embryo and in cotyledons meristems and leaves. To adequately portray the unique during early stages of germination. It was concluded that features in these species, special terms were introduced KNOX1 gene expression is tightly linked to meristematic Dev Genes Evol (2010) 220:25–40 27

Fig. 1 The unique growth patterns in Streptocarpus. a–d The rosulate Streptocarpus species. e 4–7 Acaulescent (rosulate, unifoliate) species. Streptocarpus rexii. a 1 Isocotylous stage 15 DAS. a 2 Anisocotylous e 4 The macrocotyledon develops into the “cotyledonary phyllo- stage 30 DAS. a 3 Anisocotylous seedling with first phyllomorph 65 morph” and forms a groove meristem on the petiolode. e 5 In rosulate DAS. a 4 Seedling with several leaves 90 DAS. Bars, 1 mm. b–d species, the first true leaf, or “primary phyllomorph” is formed from SEM images. b Top view and c side view of anisocotylous seedling. d the groove meristem of the macrocotyledon. e 6 Front view of a single Magnified view of c, showing the basal meristem and groove phyllomorph. e 7 TS off-center as indicated (dashed line) through a meristem. e Schematic illustration of seedling development in phyllomorph. ab abaxial (lower), ad adaxial (upper) leaf surface, b Streptocarpus (modified from Jong 1970; Jong and Burtt 1975). e 1 basal meristem, g groove meristem, Mc macrocotyledon, mc micro- Isocotylous seedling. e 2 Anisocotylous seedling. e 3 Caulescent cotyledon, p petiolode meristem, SAM shoot apical meristem 28 Dev Genes Evol (2010) 220:25–40 activities not only in the groove meristem, but also in the Reverse Transcriptase (New England Biolabs: NEB, MA, persistent activity of foci of meristematic activities in S. USA). ARP PCR products were obtained using degenerate rexii leaves (Mantegazza et al. 2007). primers designed in the MYB domain conserved between In this paper, we were specifically interested in whether the AmPHAN (A. majus), LePHAN (S. lycopersicum)and antagonistic ARP-KNOX1 expression pattern of Arabidopsis NtPHAN (Nicotiana tabacum), “sphan-a” and “sphan-d” also exists in Streptocarpus with its unusual labile vegetative (Sup 4 for all primer sequences used here), while BP was morphology. We investigated the acaulescent S. rexii and the amplified by PCR using degenerate primers based on caulescent Streptocarpus glandulosissimus to elucidate wheth- regions from KNOX1 to the homeodomain conserved er variation in expression patterns of these genes may be between BP/KNAT1 (A. thaliana), NTH20 (N. tabacum), linked to differences in their growth form. The expression of TKN1 (S. lycopersicum), and SNAP1 (A. majus,Dr.A. one STM-like KNOX1 gene in S. rexii has already been Hudson, pers. comm.), “sth1aF” and “sth5aR”.The described (Harrison et al. 2005a; Mantegazza et al. 2007). amplified DNA fragments were subcloned using the pGEM Therefore, we analyzed ARP and BP-KNOX1 gene expression T Easy Vector System (Promega, WI, USA) and sequenced to clarify their roles in these developmental processes. We through the sequencing service of RBGE, and Academia characterized leaf morphogenesis and determined the tempo- Sinica (Taiwan). ral and spatial meristematic activities in leaves, indirectly by Inverse PCR was performed to extend the ARP sequen- linking lateral vein number with leaf size, directly by ces. One microgram of genomic DNA, extracted using the epidermal cell size measurements and the localization of Qiagen DNeasy Plant Mini Kit (Qiagen, Hilden, Germany), planes of newly divided cells, and by Histone H4 expression. was digested with HindIII (NEB), then self-ligated using T4 We show that these species with simple leaf primordia and DNA ligase (NEB), and used for PCR reactions with the leaves express KNOX1 genes in the leaf. We further SrARP specific primers “stphan850F” and “stphan749R” demonstrate that both ARP and KNOX1 genes are designed here. co-expressed in the groove meristem and basal meristem in 5′-Rapid amplification of cDNA ends (RACE) was leaves in S. rexii, as well as in the SAM and basal area of the performed with the SMART RACE cDNA Amplification leaf of the caulescent S. glandulosissimus. This co- kit (Clontech, CA, USA) following the manufacturer’s expression may be linked to the extended leaf basal meristem protocol. A 3′-RACE was performed to extend the 3′-end of activity observed in both Streptocarpus growth forms. BP. After the first strand cDNA was amplified with Oligo dT-3sites Adapter primer (Takara, Otsu, Japan), the cDNA was used for PCR reactions with the 3sites Adapter primer Materials and methods (Takara) and primer “sth2dF” designed here.

Plant materials Homology analyses

Leaves for DNA and RNA extraction were harvested from Deduced amino acid sequences of the conserved MYB plants of S. rexii Lindl. and S. glandulosissimus Engl. held at regions of reported plant MYB-like proteins (Sup 2) for RBGE. Seedlings used for in situ hybridization were grown ARP, and those of conserved KNOX1, KNOX2, ELK, and from seeds, which were sterilized and grown as described homeobox domain sequences of reported KNOX1 genes before (Nishii et al. 2004). For the analysis of leaf (Sup 3) were aligned with the respective sequences development, material of A. majus, S. rexii,andS. glandu- obtained here, using Genetix version 5.0 (Genetix Ltd., losissimus was collected from plants cultivated at RBGE. New Milton, UK). Separate ARP and KNOX1 neighbor- joining trees were constructed using PAUP version 4.0b10 Morphological observations (Swofford 2002). Branch support analyses involved 100,000 bootstrap replicates. Seedlings were observed under a stereomicroscope SMZ-10 (Nikon, Tokyo, Japan), or by SEM. For SEM, the samples Gene expression analyses by RT-PCR were prepared and the growth stages defined as described previously (Nishii and Nagata 2007). ARP and BP expression in seedlings of S. rexii and S. glandulosissimus was analyzed by RT-PCR. RNA was Cloning of ARP and BP homologs from Streptocarpus extracted from different samples and treated with DNase RQ1 (Promega, WI, USA). cDNA was synchronized as Total RNA was extracted using TRIZOL (Invitrogen, CA, above and was PCR-amplified using different primer pairs USA). First strand complementary DNA (cDNA) was (Sup 4). The expression of the ACTIN gene or 18S synthesized with Oligo dT primer [d(T)18] and M-MulV ribosomal RNA (rRNA) was used as internal controls. Dev Genes Evol (2010) 220:25–40 29

Gene expression analyses by real-time PCR chloral hydrate (Nishii et al. 2004) and observed under an optic microscope BX51 (Olympus, Tokyo, Japan). Total RNA extracts were treated as above. To conduct the To pinpoint the location and duration of leaf growth in S. two-step real-time PCR, cDNA was synthesized using glandulosissimus, the epidermal cell size in different SSPIII reverse transcriptase (Invitrogen). We conduct the regions of the leaves was determined and plotted against real-time PCR using the KAPA SYBR FAST qPCR Kit different stages of leaf development. The exact cell area (Kapabiosystems, MA, USA) with gene-specific primer was measured using the graphic program NIH image (Scion sets designed here (Sup 4) using Primer Express (Applied Co. MD, USA). Biosystems Inc., CA, USA). Real-time PCR was conducted To identify dividing cells in developing leaves of S. in a BioRad RQ5 (Bio-Rad Laboratories, CA, USA) glandulosissimus, aniline blue staining for the fluorescent following the manufacturer’s protocol. The melting curve detection of β-1,3 glucan, which is contained in newly was analyzed for each experiment individually for each formed cell walls, was conducted as described before primer set. The obtained threshold cycle (Ct) values were (Nishii et al. 2004). Samples were observed under a analyzed by REST (Pfaffl et al. 2002). 18S ribosomal RNA fluorescence microscope (AX70, BX51, Olympus) using was used as internal standard. Control samples to obtain U-excitation. We observed the aniline blue stained cell relative expression levels were whole seedlings (50 days walls of epidermal and subepidermal cell layer from the after sowing, DAS) or an entire plant with several adaxial side, where no stomatal division occurs in Strepto- phyllomorphs (120 DAS) for S. rexii and a shoot apex of carpus (Noel and Van Staden 1975). a mature plant with several pairs of leaves for S. To investigate the link between the extended basal glandulosissimus. Experiments were conducted in tripli- meristem activity in S. glandulosissimus with meristem cates for each sample and repeated at least two times. genes, we analyzed the expression patterns of SglARP and Relative gene expression levels were calculated and a SglBP in leaves at different developmental stages by hypothesis test [P(H1)] performed with REST. RT-PCR and real-time PCR. To demonstrate cell division activity in the leaves of S. glandulosissimus,wealso Gene expression analyses by in situ hybridization analyzed the expression pattern of Histone H4 (H4), because H4 expression suggests a proliferating state of Digoxygenin-labeled (DIG) RNA probes for ARP and BP tissues (Fobert et al. 1994). A partial sequence of the H4 were generated using an in vitro transcription kit (Roche gene was isolated from S. glandulosissimus cDNA, with Diagnostics GmbH, Mannheim, Germany) according to the the degenerate primer pair “SiH4-MSG-F” and “SiH4- manufacturer’s protocol. ARP and BP DNA fragments were AV T-R”. To verify the homology, the deduced amino acid amplified from cDNA prepared by PCR with primers sequence of SglH4 was aligned with reported H4 genes “stphan850” and “sphan-d” for ARP and “sth1aF” and (Sup 1). RT-PCR and real-time PCR experiments were “sth5aR” for BP. The probe positions were as follows: from conducted as above. 847 to 1,041 bp for SrARP, 847–1,011 bp for SglARP, 487–874 bp for SrBP, and 517–906 bp for SglBP, respectively. Sense transcripts were used as negative Results controls. Hybridization and immunological detections were performed as described before (Mantegazza et al. 2007). Cloning of SrARP and SrBP genes Images of in situ hybridization were taken with optical microscopes Axiophot2 (Carl Zeiss Ltd., Welwyn Garden SrARP isolated from S. rexii showed a sequence similarity City, UK) and Optiphot-2 (Nikon). in the MYB domain of 88% compared to AmPHAN at the protein level and 74% identity at the nucleotide level. Assessment of leaf development and meristematic activity SglARP from S. glandulosissimus showed 98% identity at the protein level and 96% identity at the nucleotide level, We plotted the number of veins against leaf size from compared to SrARP (Fig. 2). In the protein neighbor-joining primordia initiation to final size in comparison to A. majus tree of the MYB domain, SglARP and SrARP fell next to of Plantaginaceae, a family closely related to Gesneriaceae. AmPHAN in an ARP gene clade (bootstrap support 98%), The leaves of most flowering plants produce primary lateral indicating that the Streptocarpus sequences were likely veins early in development and reach their maximum ARP homologs (Fig. 2b). number of veins long before reaching their maximum leaf SrBP isolated from S. rexii showed 86% identity at the size (Poethig and Sussex 1985a). Primary lateral veins were protein and 75% at the nucleotide level in the conserved observed by eye in leaves more than 10 mm in length. domains compared to AtBP (Fig. 3). SglBP from S. Leaves smaller than 10 mm in length were cleared with glandulosissimus was similar to SrBP by 97% at the protein 30 Dev Genes Evol (2010) 220:25–40

Fig. 2 a Alignment of Strepto- carpus ARP genes, Antirrhinum PHAN and Arabidopsis AS1, indicating the conservation of regions at the amino acid level. b Neighbor-joining tree based on deduced amino acid sequen- ces of MYB-like genes using the conserved MYB domain. Bootstrap values are shown along the branches. Branches without values received less than 50% bootstrap support. AtMYB97, FLP, AtAS1, Arabi- dopsis thaliana; SkARP, Selagi- nella kraussiana; ZmRS2, Zea mays; OSMYB4, OsRS2, Oryza sativa; NtMYBGR1, NtPHAN, Nicotiana tabacum; LePHAN, Solanum lycopersicum; AmPHAN, AmPHAN2, Antirrhinum majus; SrARP, Streptocarpus rexii; SglARP, S. glandulosissimus

and 96% at the nucleotide level. Their homology to other SrBP and SrARP in 50 DAS seedlings of S. rexii. The real- BP genes was supported by the neighbor-joining analysis, time PCR analysis confirmed that co-expression was in which SrBP and SglBP grouped with other BP-like confined to the proximal region of the macrocotyledon KNOX1 genes, including AtBP. The bootstrap value and absent from the distal region of the macrocotyledon. supporting this Eudicot BP-like KNOX1 gene clade was The vegetative developing phyllomorph of mature S. rexii 78% (Fig. 3b). plants showed SrBP and SrARP expression whereas in the flowering phyllomorph virtually no expression was detect- SrARP and SrBP are expressed predominantly able (Fig. 4b). in the proximal region of the macrocotyledon of S. rexii In situ hybridization of SrARP on anisocotylous seedlings at the flat groove meristem stage, 30 DAS, showed Shortly after germination (15 DAS) during the isocotylous expression in the proximal region of the macrocotyledon, stage, expression of SrARP and SrBP was observed in where the basal meristem is expected to form (Fig. 5a). cotyledons by RT-PCR (Fig. 4a). At this stage, no Further expression was seen in the groove meristem and in significant difference in size was observed between the the vasculature. SrARP signals were also observed in cells of two cotyledons (0.55±0.09 mm and 0.6±0.09 mm long; the bulged groove meristem, 60 DAS (Fig. 5b). At the “leaf n=12, Student’s t test, P>0.05). By 50 DAS, anisocotyly formation” stage (sensu Nishii and Nagata 2007) (65 DAS), was fully established, the length of the macrocotyledon SrARP was expressed in the developing lamina of the newly 2.65±0.27 mm and that of the microcotyledon 0.75± formed phyllomorph (Fig. 5e). At the junction of its 0.14 mm (n=14, 0

Fig. 3 a Alignment of Strepto- carpus BP and STM genes, and Arabidopsis BP and STM genes indicating the conservation of the KNOX1/2, ELK, and HOMEODOMAIN regions at the amino acid level. b Neighbor-joining tree based on deduced amino acid sequen- ces of the KNOX gene family, using the KNOX and homeodomain-related conserved region. Bootstrap values are shown along the branches. Branches without values received less than 50% bootstrap support. OSH15, Oryza sativa; rough sheath 1, Zea mays; KNAT2, KNAT6, AtSTM, AtBP, Arabidopsis thaliana; KNAP1, Malus domestica; NTH20, NTH15, Nicotiana tabacum; Tkn1, LET6, Solanum lycoper- sicum; SsSTM1, Streptocarpus saxorum; SdSTM1, S. dunnii; SrSTM1, SrSTM2, SrBP, S. rexii; SglBP, S. glandulosissimus

ARP and BP expression patterns in the caulescent leaf meristems in these species over longer periods of S. glandulosissimus development, using four different approaches. Leaves of A. majus, used here as a control species, showed no increase in As in S. rexii, SglARP and SglBP expression was observed in the number of the primary lateral veins as they grow from the proximal region of the macrocotyledon of the caulescent 10 mm to their maximum size of 80 mm indicating that this S. glandulosissimus (Fig. 5m, o). In older plants, SglARP and growth was simple expansion (Fig. 6a). Streptocarpus SglBP were co-expressed in the SAM as well as in the leaves of both species, however, showed a steadily emerging leaf primordia (leaf 1 = L1) (Fig. 5p, q, s). In L2 of increasing leaf length and number of primary lateral veins the next node, SglARP and SglBP expression was observed almost to their final leaf size, with the only difference that in the vascular tissues and throughout the lamina (Fig. 5p–v). leaves in S. rexii grew larger (Fig. 6a). This shows new leaf lamina being produced. Leaf growth and meristematic activity The leaf epidermal cell size in S. glandulosissimus during leaf development showed that initially cells in all Because the in situ results showed a very similar co- regions of S. glandulosissimus leaves were small, but those expression of ARP-KNOX1 genes in all vegetative meris- in the middle and distal region soon enlarged, from 10 mm tems, and leaf primordia, of caulescent and acaulescent leaves onwards (Fig. 6b). Cells in the proximal region Streptocarpus species, we examined and characterized the remained small for much longer, up to a leaf size of ca. 32 Dev Genes Evol (2010) 220:25–40

Fig. 4 a RT-PCR transcription patterns of SrARP and SrBP in S. rexii. pattern of SrBP (open bars) and SrARP (shaded bars). Relative Gene expression was examined by RT-PCR in isocotylous (15 DAS) expression levels as calculated by REST (±standard error bars, n=3), and anisocotylous seedlings (50 DAS), and in different regions as in various seedling tissues in relation to whole seedlings 50 DAS, and indicated, and whole seedlings were used as the control. ACTIN gene vegetatively developing and flowering phyllomorphs in relation to an was used as internal standard. In the 15 DAS seedlings, both of SrARP entire plant (120 DAS). Open bars: SrBP, shaded bars: SrARP. pMc and SrBP expressions were observed in the cotyledons. In the 50 DAS proximal region of the macrocotyledon, dMc distal region of the seedlings, ARP and BP expressions were observed in the proximal macrocotyledon, mc microcotyledon, vP vegetatively developing region of the macrocotyledon. ARP gene was very weakly detected in phyllomorph, fP flowering phyllomorph. *P(H1)<0.05 indicates the distal region of the macrocotyledon. b Real-time PCR expression significant difference to control samples

45 mm (approx. 70% final leaf size), beyond which they expressed slightly longer compared to SglARP and SglBP increased gradually and reached a similar size to cells of the (Fig. 8a). This might be due to cell division in the mid and distal region. procambium and the development of stomata that only In S. glandulosissimus cell divisions, as indicated by occur on the abaxial surface of leaves in Streptocarpus aniline blue staining, were observed throughout the (Noel and Van Staden 1975). primordium up to 4 mm size (Table 1). As develop- ment proceeded, the region with dividing cells, the arrest Fig. 5 In situ expression patterns of ARP and BP in Streptocarpus b rexii and S. glandulosissimus. a–l S. rexii. m–v S. glandulosissimus. front (Nath et al. 2003), gradually shifted towards the Left two columns: ARP. Right two columns: BP. a, c LS of proximal region (Fig. 7), and when the leaves were 20 to anisocotylous seedlings 30 DAS at the flat groove meristem stage. 29 mm long, cell division was observed only in the b, d LS of seedlings with bulged GM 60 DAS. e, g LS of seedlings proximal region. This pattern remained until leaves with newly developing first leaf 65 DAS. f, h Negative control treated ∼ with sense SrARP (f) and sense SrBP (h) probes. No blue or purple reached40to49mminsize( 70% final leaf length). signal was detected. Yellow or brown colors are background. i–l TS Beyond leaf sizes of 50 mm, cell division was no longer through petiolodes of developing phyllomorphs, at the level of the BM observed (Table 1). (i, k), and GM (j, l). m, o LS of anisocotylous seedlings 30 DAS. n RT-PCR results in S. glandulosissimus showed that both Negative control, treated with sense SglARP probe. p, q Serial TSs through the shoot apex of a mature plant. r–t Serial LSs through the SglARP and SglBP genes were expressed in 10- and shoot apex in mature plants, the level of the SAM and L2 (r, t) and L1 20-mm-long leaves. Both genes were found only in the and L2 (s). u, v Schematic illustrations of the planes of sectioning in r. proximal region of leaves 35 mm long, but not in leaves Shaded areas indicate gene expressions domains. Arrows indicate 45 mm or longer (Fig. 8a). Histone H4 expression regions of gene expression, detected as purple or blue colors. b basal meristem, g groove meristem, L1 leaf 1, L2 leaf 2, Mc macro- examined in S. glandulosissimus showed expression in all cotyledon, mc microcotyledon, pp phyllomorph primordia, SAM shoot regions of leaves up to 45 mm length and was thus apical meristem, v vascular bundle, 1p first phyllomorph Dev Genes Evol (2010) 220:25–40 33 34 Dev Genes Evol (2010) 220:25–40

The expression levels of SglBP and SglARP genes as detected by real-time PCRs were high in the proximal region of 15-mm leaves, but not in the distal region. The expression level decreased in 35-mm leaves compared to 15-mm leaves, and virtually no expression was detected in 50- or 60-mm leaves. SglH4 expression was detected in both, proximal and distal, regions of a 15-mm leaf, and proximal region of 35-mm leaves. Very little SglH4 expression was detected in the distal region of 35-mm leaves.

Discussion

KNOX1 and ARP genes in Streptocarpus

In this study, we have isolated the developmental genes ARP and BP-like KNOX1 from the acaulescent S. rexii and the caulescent S. glandulosissimus. We deduced their homology to ARP and KNOX1 genes by sequence similarity and phylogenetic analyses. Together with STM1,aSTM homolog (Harrison et al. 2005a) already cloned and characterized for acaulescent and caulescent Streptocarpus species, we are able to comparatively investigate the involvement of three genes in the meristem and leaf/phyllomorph development in Streptocarpus. Fig. 6 Leaf development in Streptocarpus and Antirrhinum. a Number of primary lateral veins plotted against leaf length for S. rexii (open circles), S. glandulosissimus (closed circles), and A. majus KNOX1 and ARP genes are involved in macrocotyledon (open squares). b Epidermal cell size on the adaxial leaf surface growth plotted against size of developing leaves of S. glandulosissimus. Cells were measured near the base (open circles), the middle (closed squares), and tip (open squares) region of leaves. Each data point is The macrocotyledon of Streptocarpus is a unique feature of the average of 10 measurements (error bars indicate SD) many Old World Gesneriaceae (Burtt 1970), and their development cannot be easily compared to cotyledons of other angiosperms. In a previous study, SrSTM1 expression seemed to be implicated in the activity of the basal meristem during macrocotyledon growth in S. rexii (Mantegazza et al.

Table 1 Prolonged cell division activity in the proximal region of S. glandulosissimus leaves

Longitudinal leaf length (mm) No. of samples Segment position

AB CD Proportion of segments showing divisions

0− 4 23 1.00 1.00 0.91 0.61 5− 9 19 1.00 1.00 0.42 0.00 10−19 18 1.00 0.72 0.00 0.00 20−29 16 0.63 0.06 0.00 0.00 30−39 15 0.60 0.00 0.00 0.00 40−49 6 0.33 0.00 0.00 0.00 50−70 8 0.00 0.00 0.00 0.00

Leaves were divided into four segments, proximal to distal. Each segment was categorized for presence and absence of cell division, as indicated by aniline blue staining, and the proportion of segments showing division calculated. A: proximal, B: proximal-middle, C: distal-middle, D: distal segment. Shading indicate absence of cell division Dev Genes Evol (2010) 220:25–40 35

Fig. 7 Development of the arrest front of cell division in Streptocarpus glandulosissimus leaves at early stages of devel- opment, as indicated by aniline blue staining. a Fluorescing newly formed cell walls after cell division under UV (arrows). b Cell division patterns of epidermal and sub-epidermal regions in leaves observed from the adaxial surface. Open circles: cell planes stained with aniline blue, closed squares: leaf contour

2007). Here, we can report that SrARP and SrBP are also undifferentiated state (Long et al. 1996; Waites et al. 1998; expressed in the basal meristem of the macrocotyledon in Byrne et al. 2000). this species (Figs. 4, 5,and9), and together with SrWUS, We demonstrate that in the Streptocarpus species another meristem gene expressed in the leaves of S. rexii analyzed here, irrespective of whether acaulescent or (Mantegazza et al. 2009), it seems that the expression of caulescent, ARP and BP-KNOX1 homologs are co- several meristem genes is spatially and temporally shifted in expressed in the SAM/groove meristem, and in the leaf/ Streptocarpus compared to Arabidopsis. However, later phyllomorph, i.e. leaf primordia and basal leaf meristem stages of leaf development, such as acquisition of adaxial/ (Figs. 4, 5, 8, and 9), and are not expressed in flowering abaxial polarity are not changed in Streptocarpus, and phyllomorphs where the basal meristematic activity had accordingly, the polarity gene FIL is expressed in S. rexii ceased (Fig. 4b). Another KNOX1 homolog, SrSTM1, has in a similar pattern as in Arabidopsis (Tononi et al. 2010). previously been shown to be expressed in S. rexii in a Although the macrocotyledon in caulescent Streptocarpus pattern very similar to SrBP here (Harrison et al. 2005a; species does not grow indeterminately as the “phyllomorph” Mantegazza et al. 2007). This is very different from the of acaulescent species (Jong 1970), co-expression of ARP expression of these genes in Arabidopsis. It may be that and KNOX1 genes was also found in the macrocotyledon of KNOX1 homologs in general show an extended expression S. glandulosissimus (Figs. 5 and 9). This suggests that there domain, extending into the leaf in Streptocarpus, but this is a common mechanism, at least for the genes analyzed remains to be tested. here, for macrocotyledon growth and subsequent meristem Previous results described the expression of KNOX1 development in Streptocarpus. proteins in the SAM of the caulescent S. saxorum (Harrison et al. 2005a), which is consistent with our results here. This KNOX-ARP expression patterns during leaf development work also showed the down-regulation of KNOX in in Streptocarpus incipient leaf primordia in the SAM, but we have not detected such down-regulation in the SAM of S. glandulo- In many angiosperms, the expression of ARP genes is sissimus. This is either due to interspecific differences, or restricted to leaf primordia, while KNOX1 genes are our limited in situ hybridization studies, because down- expressed mainly in the SAM that is maintained in an regulation of KNOX1 gene expression in incipient leaf 36 Dev Genes Evol (2010) 220:25–40

primordia may be a short-lived event that is hard to detect given the sporadic activity of Streptocarpus SAMs (J. Harrison pers. comm.). The similarity in gene expression in the present study between acaulescent and caulescent Streptocarpus species suggests that ARP and KNOX1 genes are not responsible for the differences between the SAMs of caulescent Streptocarpus and groove meristems of acaulescent Strep- tocarpus (Fig. 9, see also Mantegazza et al. 2007). However, the co-expression of ARP and KNOX1 in the SAM and leaf primordia in simple leaved species is a novel observation for angiosperms (Table 2). ARP-KNOX1 co- expression in the SAM and leaf primordia has been reported in compound leaved angiosperms, such as S. lycopersicum and ferns (Koltai and Bird 2000; Harrison et al. 2005b; Table 2). KNOX1 expression in leaf primordia was observed in simple leaved plants with complex primordia (Bharathan et al. 2002). However, Streptocarpus is simple leaved, with simple leaf primordia (Lai 2001; Barth et al. 2009). Simple leaves are ancestral in the sister orders to which Gesneriaceae belong and Solanales (including S. lycopersicum) (Bharathan et al. 2002; Wortley et al. 2005). The co-expression of ARP-KNOX1 genes could be a shared Fig. 8 a RT-PCR transcription pattern in Streptocarpus glandulosis- simus. Gene expression of SglARP and SglBP was examined in the ancestral characteristic of S. lycopersicum and Streptocar- leaves of different sizes and different regions as indicated. Histone4 pus. However, A. majus, a close relative of Streptocarpus in (SglH4) was used as indicator for cell division, and ribosomal 18S RNA Lamiales, has a mutually exclusive ARP-KNOX1 gene (r18S) as internal standard. b Real-time PCR expression pattern. Relative expression similar to A. thaliana (Waites et al. 1998). This expression level of SglARP (open bars), SglBP (shaded bars—light gray), and SglH4 (shaded bars—dark gray) as calculated by REST may indicate that the genetic pathway of ARP-KNOX1 is (±standard error bars, n=3) in S. glandulosissimus leaves of various sizes highly plastic and can alter rapidly. The variation in in relation to a shoot apex, using the 18S rRNA expression as the internal ARP-KNOX1 expression patterns across the angiosperms standard. pL proximal leaf region, dL distal leaf region. *P(H1)<0.05 phylogeny given in Table 2 supports this plasticity. indicates significant difference to control samples Whether a change in the interaction between ARP and

Fig. 9 Summary of expression pattern of ARP and KNOX1 genes in S. glandulosissimus (c), whereas in model plants with simple leaves, vegetative meristems of Streptocarpus compared to model plants. ARP such as Arabidopsis thaliana and Antirrhinum majus, ARP is and KNOX1 gene co-expression was observed in the proximal region exclusively expressed in leaf primordia, and KNOX1 in SAMs of of the macrocotyledon (a), the meristems in phyllomorphs of the seedlings (d) or mature plants (e) (Waites et al. 1998; Byrne et al. rosulate S. rexii (b), and the SAM and leaf primordia in the caulescent 2000; reviewed in Hake et al. 2004) Dev Genes Evol (2010) 220:25–40 37

Table 2 Expression profiles of ARP and KNOX1 genes/proteins in developing leaves (KNOX1 down-regulation in leaf primordia was not considered in this table) and SAM/GM of diverse plants

Species Leaf type Gene/antibody Expression in References

SAM/ leaf GM

Streptocarpus rexiia, c (a) Simple: phyllomorph SrSTM1 + (GM) + Mantegazza et al. (2007) SrBP + (GM) + This study SrARP + (GM) + Streptocarpus Simple SglBP + + This study glandulosissimusb, c (a) SglARP ++ Streptocarpus saxorumb, c (a) Simple Anti-KNOX ++Harrison et al. (2005a) Antirrhinum majusc (a) Simple AmSTM + Waites et al. (1998) PHAN (ARP)+ Nicotiana tabacumc (a) Simple NTH15 (STM homolog) + Nishimura et al. (1999) NTH20 (BP homolog) + NtPHAN + McHale and Koning (2004) Arabidopsis thalianac (b) Simple STM + Long et al. (1996) BP/KNAT1 + Lincoln et al. (1994) AS1 + Byrne et al. (2000) Lepidium oleraceumc (b) Simple: complex Anti-KNOX ++Bharathan et al. (2002) primordia Zea maysd Simple: with sheath knotted1 (KNOX1) + Smith et al. (1995) rs2 (ARP) + Timmermans et al. (1999) Tsiantis et al. (1999) Amborella trichopodae Simple Anti-KNOX + Bharathan et al. (2002) Welwitschia mirabilisf Simple Anti-KNOX ++Pham and Sinha (2003) Selaginella kraussianag Simple SkKNOX1 + Harrison et al. (2005b) SkKNOX2 + SkARP1 ++ Solanum lycopersicumc (a) Compound Let6 (STM homolog) ++Chen et al. (1997) Tkn1 (BP homolog) ++Hareven et al. (1996) LePhan (ARP) + + Koltai and Bird (2000) Cardamine hirsutec (b) Compound ChSTM + + Hay and Tsiantis (2006) ChBP ++ ChAS1 + Pisum sativumc (c) Compound Pskn1 (STM homolog) + Hofer et al. (2001) Pskn2 (BP homolog) ++Tattersall et al. (2005) Crispa (ARP)+ Osmunda regalish Compound Anti-KNOX ++Harrison et al. (2005b) anti-ARP + +

Rows of KNOX1—not shaded, those of ARP—shaded. Of the simple leaved plants, only Streptocarpus show co-expression of KNOX1 and ARP genes in both, SAM/GM and leaf. References refer to main publications of gene expression. (a)–(c) APGII, (a) Euasterid I, (b) Eurosid II, (c) Eurosid I, +: gene expressions are reported a Acaulescent rosulate b Caulescent c Dicot d Monocot e Basal angiosperm f Gymnosperm g Lycophyte h Fern 38 Dev Genes Evol (2010) 220:25–40

KNOX1 genes has occurred at the point of origin of the This appears to represent a case of convergent appropriation Gesneriaceae or within the Gesneriaceae remains to be seen. of the ARP-KNOX1 pathway to confer indeterminate cell fate to some cells in the leaf. In tomato and C. hirsuta, this Extended basal meristematic activity is associated is associated with leaflet formation, but in Streptocarpus, with KNOX1 in Streptocarpus with additional lamina area. Further studies of ARP and KNOX1 in this genus might be important to unravel the An extended meristematic activity in the basal meristem genetic mechanisms controlling cell division pattern and of acaulescent Streptocarpus species is well documented leaf organization. (Jong 1970, 1978;JongandBurtt1975), and it can be active for several years (Hilliard and Burtt 1971). Meristem activity ceases only upon the initiation of Acknowledgments The research of KN at the Royal Botanic Garden Edinburgh (RBGE) was supported by a scholarship from the inflorescences (Jong 1978). We have shown here that University of Tokyo (Japan). KN greatly acknowledges the invaluable leaves of caulescent Streptocarpus species also show technical, scientific, and financial support from the Science Division prolonged basal meristem activity in their leaves, over of RBGE. RBGE is supported by the Scottish Government Rural and growth to about 70% of their final size. In closely related Environment Research and Analysis Directorate (RERAD). Part of this study was also supported by National Science Council in Taiwan plants, such as A. majus, cell division ceases very early in [grant number NSC97-2811-B-002-027] and the top 100 research leaf development, when leaves are about 10% of their final program in National Taiwan University (NTU) [96R0044, 96R8044]. length (Nath et al. 2003). Thus, basal meristematic activity We wish to thank Prof. A. Hudson (University of Edinburgh, UK) for is a shared characteristic of acaulescent and caulescent providing sequences of SNAP1 and for his helpful comments and technical assistance. Thanks are also due to the horticultural staff, Streptocarpus. We suggest that the prolonged basal especially Ms. Sadie Barber and Mr. Steve Scott for growing plants at meristematic activity is linked to the unorthodox co- RBGE. We are grateful to Dr. M. Kawaguchi (the University of expression of ARP and KNOX1 genes in the simple Tokyo) and Dr. T.-P. Lin (NTU) for their kind support, Dr. Y.-Y. Kao Streptocarpus leaves. This idea is supported by the for providing access to the microscope facilities at NTU, Dr. K.-J. Tang and Ms. Y.-Y. Gao (TechComm, NTU) for the facilities of observation of KNOX1 expression in the basal meristem real-time PCR, and Dr. S. Okamoto (The University of Tokyo) and region of the simple leaves of Welwitschia,anotherplant Ms. Y.-J. Chen (NTU) for their helpful comments on in situ with a greatly extended basal leaf meristem (Pham and hybridization. We also acknowledge helpful suggestions from anony- Sinha 2003). Though KNOX1 and ARP gene interactions mous reviewers to improve the manuscript. in the leaf are required to form compound leaves in tomato and C. hirsuta (Kim et al. 2003a, b; Hay and Tsiantis 2006), the role of KNOX1 proteins is to act at specific leaf development stages to delay leaf maturation, enabling leaflet formation (Shani et al. 2009). In C. hirsuta, References KNOX1-STM protein also prevents the precocious exit of tissues from meristematic activities into differentiation Barth S, Geier T, Eimert K, Watillon B, Sangwan RS, Gleissberg S (2009) KNOX overexpression in transgenic pathways, thus promoting leaflet initiation (Hay and (Gesneriaceae) prolongs the activity of proximal leaf blasto- Tsiantis 2006). In both cases, KNOX1 gene expression zones and drastically alters segment fate. Planta 230:1081– in leaves seems to maintain an undifferentiated state in 1091 leaves and promotes extended cell division capability, in Barton MK, Poethig RS (1993) Formation of the shoot apical meristem in Arabidopsis thaliana: an analysis of development effect extending the meristematic phase of leaves. The in the wild type and in the shootmeristemless mutant. Develop- basal expression of KNOX1 genes in the leaves of ment 119:823–831 Streptocarpus may have the same role here, while the Bharathan G, Goliber TE, Moore C, Kessler S, Pham T, Sinha co-expression of ARP may maintain leaf identity. Func- NR (2002) Homologies in leaf form inferred from KNOX1 gene expression during development. Science 296:1858– tional analysis via transgenes and mutagenesis may 1860 address this. Bowman JL, Eshed Y (2000) Formation and maintenance of the shoot In conclusion, we have demonstrated that ARP and apical meristem. Trends Plant Sci 5:110–115 KNOX1 genes are co-expressed in the basal meristems of Burtt BL (1970) Studies in the Gesneriaceae of the Old World. XXXI. Some aspects of functional evolution. Notes R Bot Gard Edinb macrocotyledons and leaves/phyllomorphs and in meris- 30:1–10 tems (both SAM/groove meristems) of the acaulescent S. Byrne ME, Barley R, Curtis M, Arroyo JM, Dunham M, Hudson A, rexii and caulescent S. glandulosissimus. Observations of Martienssen RA (2000) Asymmetric leaves1 mediates leaf cell division pattern during leaf development suggested that patterning and stem cell function in Arabidopsis.Nature 408:967–971 acaulescent and caulescent Streptocarpus possess an ex- Byrne ME, Simorowski J, Martienssen RA (2002) ASYMMETRIC tended basal meristematic activity in leaves. This could be LEAVES1 reveals knox gene redundancy in Arabidopsis. Devel- related to ARP/KNOX1 gene co-expression in that region. opment 129:1957–1965 Dev Genes Evol (2010) 220:25–40 39

Carles CC, Fletcher JC (2003) Shoot apical meristem mainte- Long JA, Moan EI, Medford JI, Barton MK (1996) A member nance: the art of a dynamic balance. Trends Plant Sci 8:394– of the KNOTTED class of homeodomain proteins encoded by 400 the SHOOTMERISTEMLESS gene of Arabidopsis.Nature Chen JJ, Janssen BJ, Williams A, Sinha N (1997) A gene fusion 379:66–69 at a homeobox locus: alterations in leaf shape and implica- Mantegazza R, Möller M, Harrison CJ, De Luca C, Spada A (2007) tions for morphological evolution. Plant Cell 9:1289– Anisocotyly and meristem initiation in an unorthodox plant, 1304 Streptocarpus rexii (Gesneriaceae). Planta 225:653–663 Fletcher LC, Brand U, Running MP, Simon R, Meyerowitz EM (1999) Mantegazza R, Tononi P, Möller M, Spada A (2009) WUS and STM Signaling cell fate decisions by CLAVATA3 in Arabidopsis shoot homologs are linked to the expression of lateral dominance in meristems. Science 283:1911–1914 the acaulescent Streptocarpus rexii (Gesneriaceae). Planta Fobert PR, Coen ES, Murphy GJP, Doonan JH (1994) Patterns of cell 230:529–542 division revealed by transcriptional regulation of genes during Mayer KFX, Schoof H, Haecker A, Lenhard M, Jürgens G, Laux T the cell cycle in plants. EMBO J 13:616–624 (1998) Role of WUSCHEL in regulating stem cell fate in the Gifford E, Corson G (1971) The shoot apex in seed plants. Bot Rev Arabidopsis shoot meristem. Cell 95:805–815 37:143–229 McHale NA, Koning RE (2004) PHANTASTICA regulates develop- Hake S, Smith HMS, Magnani E, Holtan H, Mele G, Ramirez J (2004) ment of the adaxial mesophyll in Nicotiana leaves. Plant Cell The role of knox genes in plant development. Ann Rev Cell Dev 16:1251–1262 Biol 20:125–151 Nath U, Crawford BCW, Carpenter R, Coen E (2003) Genetic control Hareven D, Gutfinger T, Parnis A, Eshed Y, Lifschitz E (1996) The of surface curvature. Science 299:1404–1407 making of a compound leaf: genetic manipulation of leaf Nishii K, Kuwabara A, Nagata T (2004) Characterization of architecture in tomato. Cell 84:735–744 anisocotylous leaf formation in Streptocarpus wendlandii Harrison CJ, Möller M, Langdale J, Cronk QCB, Hudson A (2005a) (Gesneriaceae): significance of plant growth regulators. Ann The role of KNOX genes in the evolution of morphological Bot 94:457–467 novelty in Streptocarpus. Plant Cell 17:430–443 Nishii K, Nagata T (2007) Developmental analyses of the phyllo- Harrison CJ, Corley SB, Moylan EC, Alexander DL, Scotland RW, morph formation in the rosulate species Streptocarpus rexii Langdale JA (2005b) Independent recruitment of a conserved (Gesneriaceae). Plant Syst Evol 265:135–145 developmental mechanism during leaf evolution. Nature Nishimura A, Tamaoki M, Sato Y, Matsuoka M (1999) The expression 434:509–514 of tobacco knotted1-type class 1 homeobox genes correspond to Hay A, Tsiantis M (2006) The genetic basis for differences in leaf regions predicted by the cytohistological zonation model. Plant J form between Arabidopsis thaliana and its wild relative Card- 18:337–347 amine hirsuta. Nat Genet 38:942–947 Noel ARA, Van Staden J (1975) Phyllomorph senescence in Hilliard OM, Burtt BL (1971) Streptocarpus: an African plant study. Streptocarpus molweniensis. Ann Bot 39:921–929 University of Natal Press, Pietermaritzburg Pham T, Sinha N (2003) Role of KNOX genes in shoot development Hofer J, Gourlay C, Michael A, Ellis THN (2001) Expression of a of Welwitschia mirabilis. Int J Plant Sci 164:333–343 class 1 knotted1-like homeobox gene is down-regulated in pea Poethig RS, Sussex IM (1985a) The developmental morphology compound leaf primordia. Plant Mol Biol 45:387–398 and growth dynamics of the tobacco leaf. Planta 165:158– Imaichi R, Omura-Shimadate M, Ayano M, Kato M (2007) Develop- 169 mental morphology of the caulescent species Streptocarpus Poethig RS, Sussex IM (1985b) The cellular parameters of leaf pallidiflorus (Gesneriaceae), with implications for evolution of development in tobacco: a clonal analysis. Planta 165:170– monophylly. Int J Plant Sci 168:251–260 184 Jong K (1970) Developmental aspects of vegetative morphology Pfaffl MW, Horgan GW, Dempfle L (2002) Relative Expression of Streptocarpus. PhD Dissertation, University of Edinburgh, Software Tool (REST) for group-wise comparison and statistical UK analysis of relative expression results in real-time PCR. Nuc Jong K (1978) Phyllomorphic organisation in rosulate Streptocarpus. Acids Res 30:E36 Notes R Bot Gard Edinb 36:369–396 Reiser L, Sánchez-Baracaldo P, Hake S (2000) Knots in the family Jong K, Burtt BL (1975) The evolution of morphological novelty tree: evolutionary relationships and functions of knox homeobox exemplified in the growth patterns of some Gesneriaceae. New genes. Plant Mol Biol 42:151–166 Phytol 75:297–311 Schoof H, Lenhard M, Haecker A, Mayer KFX, Jürgens G, Laux T Kim M, McCormick S, Timmermans M, Sinha N (2003a) The (2000) The stem cell population of Arabidopsis shoot meristems expression domain of PHANTASTICA determines leaflet place- is maintained by a regulatory loop between the CLAVATA and ment in compound leaves. Nature 424:438–443 WUSCHEL genes. Cell 100:635–644 Kim M, Pham T, Hamidi A, McCormick S, Kuzoff RK, Sinha N Shani E, Burko Y, Ben-Yaakov L, Berger Y, Amsellem Z, (2003b) Reduced leaf complexity in tomato wiry mutants Goldshmidt A, Sharon E, Ori N (2009) Stage-specific suggests a role for PHAN and KNOX genes in generating regulation of Solanum lycopersicum leaf maturation by class 1 compound leaves. Development 130:4405–4415 KNOTTED1-LIKE HOMEOBOX proteins. Plant Cell 21:3078– Koltai H, Bird DM (2000) Epistatic repression of PHANTASTICA and 3092 class 1 KNOTTED genes is uncoupled in tomato. Plant J 22:455– Smith LG, Jackson D, Hake S (1995) Expression of Knotted1 marks 459 shoot meristem formation during maize empryogenesis. Dev Lai TH (2001) Comparative morphological and anatomical analysis of Genet 16:344–348 diverse Madagascan Streptocarpus morphs and Saintpaulia. MSc Swofford DL (2002) PAUP*: phylogenetic analysis using parsimo- Dissertation, University of Edinburgh, UK ny (*and other methods), version 4. Sinauer Associates, Lincoln C, Long J, Yamaguchi J, Serikawa K, Hake S (1994) A Sunderland Knotted1-like homeobox gene in Arabidopsis is expressed in Tattersall AD, Turner L, Knox MR, Ambrose MJ, Ellis THN, Hofer the vegetative meristem and dramatically alters leaf morphol- JMI (2005) The mutant crispa reveals multiple roles for ogy when overexpressed in transgenic plants. Plant Cell PHANTASTICA in pea compound leaf development. Plant Cell 6:1859–1876 17:1046–1060 40 Dev Genes Evol (2010) 220:25–40

Timmermans MCP, Hudson A, Becraft PW, Nelson T (1999) ROUGH Vollbrecht E, Veit B, Sinha N, Hake S (1991) The developmental gene SHEATH2: a Myb protein that represses knox homeobox genes in Knotted-1 is a member of a maize homeobox gene family. Nature maize lateral organ primordia. Science 284:151–153 350:241–243 Tononi P, Moeller M, Bencivenga S, Spada A (2010) GRAMINIFO- Waites R, Selvadurai HRN, Oliver IR, Hudson A (1998) The LIA homolog expression in Streptocarpus rexii is associated with PHANTASTICA gene encodes a MYB transcription factor the basal meristems in phyllomorphs, a morphological novelty in involved in growth and dorsoventrality of lateral organs in Gesneriaceae. Evol Dev 12:61–73 Antirrhinum. Cell 93:779–789 Tsiantis M, Schneeberger R, Golz JF, Freeling M, Langdale JA (1999) Wortley AH, Rudall PJ, Harris DJ, Scotland RW (2005) How much The maize rough sheath2 gene and leaf development programs in data are needed to resolve a difficult phylogeny?: case study in monocot and dicot plants. Science 284:154–156 Lamiales. Syst Biol 54:697–709