Cardiocrinum DNA barcodes 

In order to shed some light on the actual identity of Cardiocrinum cathayanum (hort.) and the relationship, if any, between Cardiocrinum cordatum and Cardiocrinum cathayanum, I have had DNA barcodes produced for a number of taxa in the collection.

 The barcode process 

The system of DNA barcodes for organisms is based on the concept of identitifying one or more genes which can be considered typical of the genus and species and using the sequencing of these genes to identify the organism. _[1]

For plants, the chosen genes are the matK and rbcL genes in the chloroplast.

In general, it's not considered possible to identify taxa beyond the Species level _[2] although claims have been made to differentiate C. giganteum giganteum from C. giganteum yunnanense _[3]. However, this process also used the trnL-F gene sequence.

The criteria for identifying a taxa are arbitary but the following definitions, based on the work done by Meier, _[5] are in frequent use. _[4]

• Best close match, (BCM). Distance < 1%, (97% correct ID)
• Close match, (CM). 3% > Distance ≥ 1%, (60% correct ID)

 Reference Cardiocrinum barcode sequences 

The BOLD_[7] barcode database has nine entries for Cardiocrinum matK and rbcL genes.

This covers all three species and, in fact, all varieties except Cardiocrinum cordatum v. glehnii, but unfortunately there isn't rbcL data for C. cathayanum. This gap was filled by the good offices of the Royal Botanic Garden at Kew who extracted DNA from a descendant of Kew 2007-297_[8], which was then sent for barcode sequencing along with the collection samples.

This gives the list of reference sequences shown below, (Plant, gene, source)

1. Cardiocrinum cathayanum, matK, ex Kew 2007-297
2. Cardiocrinum cathayanum, matK, BOLD, GBVU2911-13
3. Cardiocrinum cordatum, matK, BOLD, GBVC3452-11
4. Cardiocrinum cordatum, matK, BOLD, GBVC3453-11
5. Cardiocrinum giganteum, matK, BOLD, GBVU2910-13
6. Cardiocrinum giganteum var. yunnanense, matK, BOLD, GBVT2686-13
7. Cardiocrinum cathayanum, rbcL, ex Kew 2007-297 ‡
8. Cardiocrinum cordatum, rbcL, BOLD, GBVC3454-11
9. Cardiocrinum giganteum, rbcL, BOLD, GBVC3455-11
10. Cardiocrinum giganteum var. yunnanense, rbcL, BOLD, GBVT2775-13
11. Cardiocrinum giganteum var. yunnanense, rbcL, BOLD, GBVC3456-11

‡ as of January, 2015 the rbcL gene has only been sequenced for 538bp.

Comparisons were made using one or both of the online BOLD Plant sequences IDS_[9] and the Align Sequences Nucleotide BLAST_[10].

Comparing all possible combinations of the above gives the tables below.

The values in the boxes are the p-distance. _[11].

What can be understood from these comparisons is,

• there is very little variation between the samples: all comparisons fall with the BCM criteria,
• Whilst all variations fall within the BCM area, the largest variations are within matK, as would be expected as this is described as offering the greatest discriminatory power._[2]
• there can be variations between reportedly identical taxa, cf. samples 3/4 where d = 0.2, so this, which can be written as 0.1±0.1, needs to be used as a lower limit for the distance values._†

 Collection Cardiocrinum barcode sequences 

The plants from the collection chosen for the first set of sequences were, [Plant, Collection code, (Initial acquisition code)],

1. Cardiocrinum cathayanum (hort.), 55o2, (7p0)
2. Cardiocrinum cathayanum (hort.), 95o1, (13p0)
3. Cardiocrinum cathayanum (hort.), 110p0, (110p0)
4. Cardiocrinum cordatum, 108o3, (4p0)
5. Cardiocrinum cordatum, 15p0, (15p0)
6. Cardiocrinum cordatum glehnii, 98o2, (6p0)

These were chosen to investigate the two questions referred to at the start: the nature of the plants sold to me as C. cathayanum, whose quoted identity I have reason to doubt and which I refer to as C. cathayanum (hort.), and any relationship there might be between C. cathayanum and C. cordatum.

Samples were sent to the DNA Sequencing Facility of the Natural History Museum who carried out the extraction and sequencing of the DNA of the barcode genes._[12]

As the trnL gene is mentioned as a possible species identifier_[13], I also arranged to have it sequenced for the collection samples and for the material provided by Kew for C. cathayanum.

 The results for C. cathayanum (hort.) 

The tables below show the results of the comparisons with the samples themselves and with the reference samples.

matK genes

rbcL genes

trnL genes

 Discussion 

In view of the fact that all differences fall within the BCM range, no certain identification can be made on the basis of these DNA Barcodes and morphology remains the best indicator of taxonomy.

However, there are variations greater than 0.2 and these can be used to suggest the following,

• 55o2 and 95o1 are very similar plants. cf. matK, rbnL and trnL self-comparisons,
• 110p0 shows some differences but which, for the rbnL and trnL genes, are within the 0.2 limit,
• it's highly unlikely that any of the three is C. cathayanum, cf. the matK and trnL comparisons with the Kew C. cathayanum references shown on the graph below.
• It seems most likely that they are what their morpholgy suggests, varieties of C. giganteum, cf. matK comparisons 5 & 6, (yellow data), on the graph.

   Data

I shall continue to label then as C. cathayanum (hort.), but do not intend to change my current taxonometric key which identifies them as C. giganteum yunnanense.

 C. cathayanum & C. cordatum 

My uncertainty about C. cathayanum is caused by the fact that the only difference I have found in the descriptions of cathayanum and cordatum is in the way the leaves are distributed along the flowering stem.

cathayanum: "Leaves absent in proximal part of stem except basal ones, crowded in middle part, laxly spirally alternate in distal part" _[14]

cordatum: "leaves loosely whorled on lower half of the stem" _[15]

The problem with these descriptions is that C. cordatum can have the distribution described for cathayanum as shown in the pictures below.

Therefore, the possibility exists that C. cathayanum & C. cordatum are not separate species but varieties of one species and the DNA barcodes were produced to investigate this possibility.

The tables below show the results of the comparisons with the samples themselves and with reference samples of C. cathayanum and C. cordatum.

matK genes

rbcL genes

trnL genes

 Discussion 

Little can be deduced in the absence of a matK sequence for 98o2.

• The three samples seem to be identical although two were C. cordatum and one was C. cordatum glehnii.
• The trnL gene comparisons show that C. cathayanum is very unlikely to be identical to cordatum samples.
• However, the matK comparisions suggest that the difference between cathayanum and cordatum is not as great as between giganteum and cathayanum.

   Data

The results tend to support the conjecture that C. cathayanum and cordatum are closely related and I shall try to get barcodes from other strains of C. cordatum in future years._†

In the meantime I've altered my taxonometric key as I no longer believe that the leaf distribution can be used to distinguish C. cathayanum from cordatum. For the time being, I use the unsatisfactory, hopefully temporary criterion of 'collected location: China' to distinguish them and would be very grateful to anyone who can offer a better!

† To do list,

• Find more matK sequences for Cardiocrinum with a view to refining the lower limit for distance variation.
• Get more sequences of C. cordatum types and, if possible, genuine cathayanum to make the analysis of their differences more significant.
• Extend the sequencing to examine the C. giganteum types to see if C. giganteum yunnanense should be considered as a variety or a form and to refine the position of C. gig. alba and C. gig. yunnanense 'Big and Pink'.