Distributions and Collection Sites of Australian Hybanthus.
Distributions of each species are shown in gray for the first species listed for each map and in black for the second. Collections are coded with letters and precise locations indicated with lines. The original location of collections of dried material from herbarium specimens is indicated with an asterisk. Hybanthus are seen to inhabit all but the most arid deserts in Australia.

Scatterplot of H. floribundus Peptide Masses and Retention Times.
Individual species are indicated with different symbols. Masses range from ∼3000 to 3500 D and retention times from 10 to 20 min. Horizontal clusters of peaks reflect possible common peptides. Overlapping peaks have been horizontally dispersed for clarity.

Mass and Retention Time Distributions for Australian Hybanthus Cyclotides.
Peptides vary in mass from 2800 to 3700 D and in retention time from 10 to 20 min. Also shown is the mass distribution for published cyclotides (gray bars).

Diagram of Potential Common Peptides Shared between Different Species of Australian Hybanthus.
Individual species are denoted by ovals or rectangles (subsections Variabiles and Suffruticosi, respectively). Peptides are seen as potentially common to two collections if they fall within 1 D and 1 min. Cases of two or more individual plants of two species sharing a potential common peptide are shown with a solid line. Instances of one or more individuals of two species sharing a potential common peptide are shown with a dashed line. The single peptide of H. stellarioides only had potential overlap with a peptide from H. vernonii with a looser mass difference restraint of 6 D.

Representative Sequences of Cyclotides from the Violaceae and Rubiaceae.
All peptides are head-to-tail cyclized, and the sequences are written starting from the presumed N-terminal cleavage point in the linear precursors (Jennings et al., 2001; Dutton et al., 2004). Sequences are aligned based on the conserved Cys residues, which are numbered in Roman numerals according to the order in which they appear in the linear precursor. Loops corresponding to the backbone segments between Cys residues are numbered for reference. Loop 6 is formed at the cyclization point of the linear precursor. References for the listed cyclotides are as follows: cycloviolin C (Hallock et al., 2000), varv A (Claeson et al., 1998), Hypa A (Broussalis et al., 2001), cyclopsychotride (Witherup et al., 1994), circulin A (Derua et al., 1996), palicourein (Bokesch et al., 2001), and kalata B1 (Saether et al., 1995). A ribbon diagram of the cyclotide framework based on the structure of kalata B1 (PDB ID 1NB1) with the cystine knot core rendered and numbering of loops is shown in the middle of the figure. At the bottom of the figure are two sequence logos (Crooks et al., 2004) to graphically illustrate sequence conservation within known cyclotides. The size of the one-letter amino acid code reflects the relative abundance of that residue in known cyclotides. Separate logos are shown for Möbius (top) and bracelet (bottom) subfamilies of cyclotides (Craik et al., 1999).

Sample LC-MS Traces of Selected Collections.
Total ion count is graphed against retention time. Variations in peptide concentration and diversity are seen, and the dried herbarium specimen of H. floribundus subsp chloroxanthus is shown to give comparable data quality to fresh material. The peaks of Hyfl A-C sequenced in this study are indicated in the trace of H. floribundus subsp floribundus with corresponding letters.

Scatterplots of Hybanthus Peptide Masses and Retention Times.
Individual species and collection locations are indicated with different symbols in each plot. Collections of individual plants from one location are differentiated by shading. Masses range from 2800 to 3700 D and retention times from 10 to 20 min. Horizontal clusters of peaks reflect possible common peptides. Overlapping peaks have been horizontally dispersed for clarity.

Schematic Representation of the General Organization of Cyclotide Precursors.
Each precursor protein consists of an ER signal, a pro-region, one to three copies of a mature cyclotide domain preceded by an NTR sequence within the pro-region, and a short tail region. The latter typically comprises four to seven hydrophobic amino acids. The various elements are drawn roughly to scale, with the ER signal typically 22 amino acids, the propeptide (including NTR) typically 40 to 60 amino acids, and the cyclotide domain typically 30 amino acids. A selection of precursor sequences is shown to illustrate the strategy used for designing primers to detect cyclotides by RT-PCR. For clarity, only the ER signal sequences and part of the cyclotide sequences are shown. Oak1 is the precursor for kalata B1 (Jennings et al., 2001) and represents cyclotide precursors from the Rubiaceae family. Voc1, Voc2 (Dutton et al., 2004), and Tricyclon (Mulvenna et al., 2005) from the Violaceae family produce the mature cyclotides cycloviolacin O8, cycloviolacin O11, and tricyclon A, respectively. In this study, a conserved sequence element, AAFALPA (black box), was identified within the ER signal of the Violaceae precursors and used as a new primer binding site. Other conserved elements used as primer targets were the CGETC and CVWIPC sequences within the cyclotide domain (light and dark gray boxes, respectively). The new Hybanthus cyclotides that were identified using these primers are indicated after each primer binding site (arrows). Hyfl F was identified using a primer based on the SISC sequence in loop 1 of Hyfl A (data not shown in this figure). The partial ER signal and the most conserved part of the NTR region are shown for the seven cyclotides that were identified using the AAFALPA primer. Secondary structure prediction of the pro-region by the program PROFsec identified a short β-sheet and an α-helix within the NTR (illustrated underneath the sequences) and a further two helical segments in the pro- region upstream of the NTR (data not shown).