Content uploaded by Robert W Pearcy
Author content
All content in this area was uploaded by Robert W Pearcy
Content may be subject to copyright.
THE TAXONOMIC DISTRIBUTION OF C
4
PHOTOSYNTHESIS IN
AMARANTHACEAE SENSU STRICTO
1
ROWAN F. SAGE,
2,3
TAMMY L. SAGE,
3
ROBERT W. PEARCY,
4
AND THOMAS BORSCH
5,6
3
Department of Ecology and Evolutionary Biology, University of Toronto, 25 Willcocks Street, Toronto, Ontario M5S 3B2
Canada;
4
Section of Evolution and Ecology, University of California, Davis, California 95616 USA; and
5
Nees-Institute for
Biodiversity of Plants, Universita¨t Bonn, Meckenheimer Allee 170, 53115 Bonn, Germany
C
4
photosynthesis evolved multiple times in the Amaranthaceae s.s., but the C
4
evolutionary lineages are unclear because the
photosynthetic pathway is unknown for most species of the family. To clarify the distribution of C
4
photosynthesis in the
Amaranthaceae, we determined carbon isotope ratios of 607 species and mapped these onto a phylogeny determined from matK/
trnK sequences. Approximately 28% of the Amaranthaceae species use the C
4
pathway. C
4
species occur in 10 genera—Aerva,
Amaranthus, Blutaparon,Alternanthera,Froelichia, Lithophila,Guilleminea,Gomphrena, Gossypianthus, and Tidestromia.
Aerva, Alternanthera, and Gomphrena contain both C
3
and C
4
species. In Aerva, 25% of the sampled species are C
4
.In
Alternanthera, 19.5% are C
4
, while 89% of the Gomphrena species are C
4
. Integration of isotope and matK/trnK data indicated C
4
photosynthesis evolved five times in the Amaranthaceae, specifically in Aerva, Alternanthera,Amaranthus,Tidestromia, and a
lineage containing Froelichia, Blutaparon, Guilleminea,Gomphrena pro parte, and Lithophila. Aerva and Gomphrena are both
polyphyletic with C
3
and C
4
species belonging to distinct clades. Alternanthera appears to be monophyletic with C
4
photosynthesis originating in a terminal sublineage of procumbent herbs. Alpine C
4
species were also identified in Alternanthera,
Amaranthus, and Gomphrena, including one species (Gomphrena meyeniana) from 4600 m a.s.l.
Key words: alpine; Alternanthera; Amaranthaceae; C
4
photosynthesis; Gomphrena.
Currently, 19 families of higher plants are known to contain
species expressing the C
4
photosynthetic pathway. In each
family, the C
4
pathway arose independently, producing
approximately 50 distinct evolutionary lineages (Sage, 2004;
Muhaidat et al., 2007). Sixteen of these families are eudicots.
The eudicot clade with the largest number of C
4
species is the
Amaranthaceae/Chenopodiaceae alliance, with half of the
approximately 1400 eudicot C
4
species. The monophyly of
the alliance has been well established, although relationships
among its major lineages are not yet clear (Manhart and Rettig,
1994; Downie et al., 1997; Cue´noud et al., 2002; Kadereit et
al., 2003; Mu¨ller and Borsch, 2005a). The major lineages are
the family Amaranthaceae (as in the circumscription of Schinz,
1893), the subfamily Polycnemoideae (currently classified
within Chenopodiaceae; Ulbrich et al., 1934; Ku¨ hn et al.,
1993), and the chenopod lineages Betoideae, Chenopodioi-
deae-Corispermoideae, and Salicornioideae-Salsoloideae-Suae-
doideae (Kadereit et al., 2003; Mu¨ller and Borsch, 2005a). It
has been proposed that the name Amaranthaceae should be
applied to all species in the alliance (Amaranthaceae s.l.,
Baillon, 1887; Malligson, 1922; APG, 2003). However,
because recent molecular studies cannot exclude the mono-
phyly of Amaranthaceae s.s. and Chenopodiaceae, these family
names are often maintained (Kadereit et al., 2003; Welsh et al.,
2003; Mu¨ller and Borsch, 2005a; Kapralov et al., 2006). The
chenopodiaceous lineages comprise the largest number of C
4
eudicot species (about 500; Sage et al., 1999). The Amar-
anthaceae s.s. contains the next highest number, previously
estimated to be around 250 C
4
species (Sage et al., 1999).
Comprehensive surveys of the occurrence of C
4
photosyn-
thesis in eudicot families have focused on the Chenopodiaceae,
most notably the tribes Salsoleae and Suaedeae (Akhani et al.,
1997; Pyankov et al., 1997, 2001a, b; Schu¨tze et al., 2003).In
these tribes, there is strong evidence for multiple evolutionary
origins of the C
4
pathway, as indicated by variation in leaf
anatomy, biochemical metabolism, and the occurrence of the
C
4
species on separate branches of molecular-based phyloge-
netic trees (Pyankov et al., 2001a, b; Kadereit et al., 2003). By
contrast, little work on the distribution of C
4
photosynthesis in
the Amaranthaceae s.s. has been reported since initial surveys
were conducted three decades ago (Downton, 1975; Ragha-
vendra and Das, 1978; Ziegler et al., 1981). These initial
surveys were incomplete because they mainly examined
common species and often reported only those with C
4
photosynthesis. Without a clear picture of whether species
are C
3
,C
4
, or intermediate between C
3
and C
4
, it is impossible
to assess where the transition from C
3
to C
4
photosynthesis
occurs. Multiple origins appear likely because of the
occurrence of known C
4
species in different tribes that also
contain C
3
species (Sage, 2004; Kadereit et al., 2003). In the
1
Manuscript received 19 January 2007; revision accepted 9 October
2007.
The authors thank M. Nee and N. Tarnowsky of the New York
Botanical Garden, J. Solomon of the Missouri Botanical Garden, S. Atkins
and K. Vollesen of Kew Gardens, and E. Wood of the Arnold Arboretum
and Grey Herbarium for assisting in the access to herbarium materials. S.
Clemants of the Brooklyn Botanical Garden kindly provided pre-published
species lists of many Amaranthaceae genera. M. Frohlich of the British
Museum of Natural History provided specimens from the British Museum
Herbarium. L. Craven of the Australian National Herbarium and D.
Kubien of the University of New Brunswick provided specimens of
Australian Gomphrena. The authors are grateful to K. Mu¨ller (Bonn) for
providing an unpublished trnK/matK sequence for Amaranthus praeter-
missus, and to G. Kadereit (Mainz) and K. Wilhelm (Oldenburg) for
valuable comments on an earlier version of the manuscript. Funding was
provided by grants BO 1815/1–1 and 1–3 and a Heisenberg fellowship
from the Deutsche Forschungsgemeinschaft (DFG) to T.B. and grant
OGP-0154273 from the Natural Science and Engineering Council of
Canada to R.F.S.
2
Author for correspondence (e-mail: R.sage@utoronto.ca)
6
Current address: Plant Biodiversity and Evolution Group and
Botanical Garden, Carl von Ossietzky Universita¨ t Oldenburg, 26111
Oldenburg, Germany
1992
American Journal of Botany 94(12): 1992–2003. 2007.
Amaranthaceae, three independent origins of C
4
photosynthesis
were inferred from a phylogeny based on rbcL sequences,
(Kadereit et al., 2003). More recently, phylogenetic analyses
using additional taxa and matK/trnK sequences were able to
better resolve relationships within the Amaranthaceae s.s. and
revealed the presence of previously unknown clades (Mu¨ ller
and Borsch, 2005a, b). With this matK/trnK data, we can now
more precisely map C
4
species onto the Amaranthaceae tree.
To do this, however, a detailed survey of the presence of C
3
and C
4
photosynthesis in the genera of the Amaranthaceae is
required.
Here, we present results of a comprehensive survey of the
Amaranthaceae s.s. for C
3
and C
4
photosynthesis. Our
approach was to sample herbarium specimens for stable carbon
isotope ratio (d
13
C). C
4
plants have d
13
C values between
10%and 15%, while C
3
species generally have d
13
C
values between 20%and 33%(Ehleringer et al., 1997).
Because C
4
-like carbon isotope ratios also occur in CAM
species, we also examined the leaves for the presence of
succulence and Kranz anatomy to rule out the possibility that
the isotopic value was due to CAM photosynthesis. In total, we
examined over 600 of the 900 or so species in the
Amaranthaceae s.s. (Townsend, 1993). Moreover, using
matK/trnK sequence data, we inferred a phylogenetic tree with
taxon sampling extended to better match the C
4
species in our
isotope survey.
MATERIALS AND METHODS
Sampling strategy and material—Samples were collected from herbarium
specimens in the collections of the Australian National Herbarium, Canberra,
Australia (CANB); the Arnold Arboretum (AH) and Grey Herbarium (GH) at
Harvard University, Cambridge, Massachusetts, USA; the herbarium of the
Charles Darwin University at Palmerston, NT Australia (DNA-NT); the
Missouri Botanical Garden, St. Louis, Missouri, USA (MO); the New York
Botanical Garden (NY) New York, USA; and the Royal Botanical Gardens,
Kew, Richmond, UK (K). Additional samples were investigated from the
Amaranthaceae research collection at the herbaria of the Nees-Institute,
University of Bonn (BONN) and the University of Texas, Austin (TEX). A list
of the specimens sampled is provided in Appendixes S1–S3 (see Supplemental
Data with the online version of this article).
Kranz anatomy and isotope analysis—Sampling consisted of first
illuminating the back of the herbarium sheet with a microscope lamp to
highlight the leaf venation. Venation was examined with a 203hand lens, and
if the leaf had large, dark veins with small aereoles, it was recorded as having
Kranz anatomy. Leaves with small, clear veins with relatively large interveinal
areas were recorded as lacking Kranz anatomy. After venation was recorded, a
small piece (,10 mg) of leaf, stem, or root tissue was removed from the
herbarium sheet, or if available, from a packet of loose material attached to the
sheet. Samples were stored in microcentrifuge tubes until placed in a tin sample
cup (#D1008, Elemental Microanalysis, Okehampton, UK) and sent to the
University of California, Davis mass spectrometer facility for isotopic analysis
(http://stableisotopefacility.ucdavis.edu). Samples were analyzed with a mass
spectrometer using a Pee Dee Belamnite limestone standard. Two or more
different herbarium specimens were examined when available. If sampling was
not permitted (as with type specimens or sheets with little material), the
venation pattern was examined and reported if it was obviously Kranz or non-
Kranz.
Molecular phylogenetics—The distribution of C
4
photosynthesis was
plotted on a tree obtained from parsimony analyses of a combined matK/trnK
data set. Most sequences were taken from Mu¨ller and Borsch (2005a, b), but the
sequences for Pedersenia and Xerosiphon were from the lab of T. Borsch
(unpublished data). The sequences for Aerva sanguinolenta, Alternanthera
altacruzensis, Al. flavescens,Al.microphylla,Amaranthus asplundii, and Am.
viridis were generated for this study using methods described in Mu¨ller and
Borsch (2005a). See Appendix S4 (with Supplemental Data in online version of
this article) for specimen information. Alignment of length variable sequences
followed the rules described in Lo¨hne and Borsch (2005) using PhyDE (Mu¨ller
et al., 2006). From the overall matrix of 2856 positions, five mutational
hotspots had to be excluded (H1 ¼position [pos.] 543–577, H2 ¼639–727, H3
¼899–929, H4 ¼1536–1538, H5 ¼2603–2628) that corresponded to hotspots
found in earlier analyses. The resulting matrix comprised 2672 positions from
which 492 were variable plus 716 that were variable and parsimony
informative. Three inversions were detected (pos. 166–185 of the 50trnK
intron in Al. caracasana and Al. pungens; pos. 249–257 of the matK CDS in
Pfaffia fruticulosa; pos. 600–602 in many taxa, ¼HS 4). The first two indels
were reverse-complemented before analysis following Lo¨ hne and Borsch
(2005), whereas the third was excluded because of its oscillating nature (Borsch
and Quandt, in press). Indels were coded in a separate, binary matrix using the
simple indel-coding approach as implemented in the program SeqState (Mu¨ller,
2004). The sequence data set (trnK group II intron and matK gene) and the
indel matrix were combined for all analyses because earlier studies had shown
that individual trees did not conflict in Amaranthaceae and that indels are
phylogenetically informative (Mu¨ ller and Borsch, 2005a). Maximum parsimo-
ny searches were conducted with the ratchet algorithm (Nixon, 1999) by
creating command files with the program PRAP (Mu¨ ller, 2004) and executing
them in the program PAUP* (Swofford, 1998). Ratchet settings were 10
random addition cycles in 200 iterations each, with the weight of 25%
perturbed characters ¼2. Node support was calculated via jackknifing (10 000
replicates, 36.8% character deletion, heuristic searches holding only one tree).
The distribution of C
4
photosynthesis was then plotted on tree number one
obtained by parsimony using the program Mesquite 1.12 (Maddison and
Maddison, 2006).
Taxonomic treatment—Circumscription of genera largely follows Town-
send (1993) with modifications indicated by recent molecular phylogenies
(Mu¨ller and Borsch, 2005a, b). Our treatment also reflects results of
phylogenetic work in progress such as the study of Iresine and relatives (T.
Borsch lab, unpublished data). To obtain an up-to-date species nomenclature,
we used several recent checklists and floras, including the Cata´ logo de las
Plantas Vasculares de la Repu´blica Argentina (Pedersen, 1999), Flora of
Ecuador (Eliasson, 1987), Cata´ logo de las Plantas Vasculares de Peru
(Borsch, 1993), Cata´logo de las Plantas Vasculares de Bolivia (Borsch et al.,
in press), Flora of East Tropical Africa (Townsend, 1985), Flora of
Madagascar (Cavaco, 1954), Flora of North America (Robertson and
Clemants, 2003), the conspectus of Australian Gomphrena (Palmer, 1998),
and a list of accepted names and synonyms prepared by Steve Clemants
(Brooklyn Botanical Garden, unpublished data). Any species name in dispute
was included in our data tables if numerous recent floras used the name.
Specimens from species no longer recognized as valid are listed with their
accepted synonyms in Appendices S1–S3 (see Supplemental Data with the
online version of this article).
RESULTS
We examined 75 of the 77 recognized genera of the
Amaranthaceae for carbon isotope ratio, often assaying the
majority if not all species within a genus. The genera not
examined because of a lack of available material were the
monotypic genera Pseudosericocoma and Hebanthodes Peder-
sen (close to Hebanthe and probably C
3
). Genera examined here
that have been recognized since Townsend (1993) include
Hebanthe, which forms a separate lineage within Gomphrenoi-
deae (Mu¨ller and Borsch, 2005a; I. Sa´nchez-del Pino and T.
Motley, New York Botanical Garden, with T. Borsch,
unpublished data); Lecosia, a Brazilian genus recognized by
Pedersen (2000); Pedersenia, which constitutes an isolated
lineage in Gomphrenoideae (T. Borsch lab, unpublished data; I.
Sa´nchez-del Pino, T. Motley, and T. Borsch, unpublished data.);
and Quaternella and Xerosiphon (Pedersen, 1990). In total, we
examined 607 species in the 77 genera listed in Table 1.
Of the 607 species examined, 233 (38.5%) were identified as
December 2007] SAGEETAL.—C
4
PHOTOSYNTHESIS IN THE AMARANTHACEAE 1993
TABLE 1. Photosynthetic pathways of plant species in Amaranthaceae s.s. Species and genera with C
4
species are indicated in bold. Data are d
13
C values
(followed by sample size where N .1). Genera values are means of the species values. When isotope analysis was not possible, the presence of Kranz
(K) or non-Kranz anatomy (NK) is indicated. See Appendix S1 (with Supplemental Data in online version of this article) for complete list of samples
and voucher information. Treatments of speciose genera such as Achyranthes,Aerva,Cyathula,Pandiaka, and Psilotrichopsis are preliminary because
many of these genera are not monophyletic, and species-level taxonomic revisions are lacking.
Taxon d
13
C%(N, if .1)
1) Achyranthes 28.7
1. ancistrophora 31.2
2. arborescens 30.2 (3)
3. aspera 28.5 (3)
4. bidentata 32.7
5. diandra NK
6. fauriei 28.2 (3)
7. japonica 30.2 (3)
8. leptostachya 30.5
9. longifolia 30.2 (2)
10. mangarevica 28.2
11. margaretarum 28.9
12. ogatai 28.6
13. robynsi 26.2
14. sessilis 27.3 (2)
15. splendens 24.9 (2)
16. talbotii 30.9 (2)
17. velutina 23.2
2) Achyropsis 26.3
1. avicularis NK
2. fruticulosa 26.7
3. gracilis 24.4
4. laniceps 28.6
5. leptostachya 25.6
3) Aerva (25%C
4
)13.0 for C
4
and
26.5 for C
3
1. artemisioides 24.2 (4)
2. congesta 26.1 (3)
3. coriacea 26.0 (2)
4. glabrata 25.0 (2)
5. javanica 13.2 (3)
6. japonica 13.8 (2)
7. lanata 25.8 (7)
8. leucura 26.9 (5)
9. madagassica 26.5 (2)
10. microphylla 28.6 (3)
11. pseudotomentosa 11.9 (2)
12. revoluta 26.7 (4)
13. ruspolli 13.1
14. sanguinolenta 29.2 (5)
15. triangularfolia 26.4 (3)
16. wightii 26.9
4) Allmania nodiflora 26.0 (5)
5) Allmaniopsis fruticulosa 27.8 (2)
6) Amaranthus 12.9
1. acanthochiton 12.3 (3)
2. acutilobus 13.3 (3)
3. albus 12.7 (4)
4. arenicola 13.4 (2)
5. asplundii 12.9 (3)
6. atropurpureus 12.4
7. australis 15.1 (3)
8. blitoides 12.5 (3)
9. blitum 13.6 (3)
10. brandegei 14.8
11. buchtienianus 12.6 (2)
12. californicus 12.6 (3)
13. cannabinus 13.7 (3)
14. cardenasianus 14.7 (2)
15. caudatus 12.3 (3)
16. celosiodes 12.1 (3)
17. crassipes 13.3 (4)
18. crispus 13.1 (2)
19. cruentus 12.4 (3)
20. deflexus 14.1 (4)
Taxon d
13
C%(N, if .1)
21. dubius 12.8 (2)
22. emarginalis 12.2 (2)
23. fimbriatus 12.8 (5)
24. floridanus 12.5 (2)
25. graecizans 13.6 (3)
26. greggii 13.3 (3)
27. hybridus 13.9 (4)
28. hypochondriachus 12.2
29. interuptus 14.2
30. kloosianus 12.0
31. leptostachyus 12.7
32. macrocarpus 11.3 (2)
33. mitchellii 11.3
34. muricatus 12.7 (2)
35. myrianthus 13.4 (2)
36. obcordatus 11.4 (2)
37. pallidiflorus 11.8
38. palmeri 12.6 (4)
39. patulus 12.7 (2)
40. peruvianus 13.8 (2)
41. persimilis 13.6 (2)
42. polygonoides 13.1 (3)
43. powellii 13.6 (3)
44. praetermissus 13.3
45. pringlei 12.6 (2)
46. pumilis 11.8 (2)
47. quintensis 12.1 (2)
48. retroflexus 13.3 (2)
49. rudis 13.7 (3)
50. scariosus 13.8 (2)
51. sclerantoides 13.0 (3)
52. spinosus 12.6 (3)
53. squamulatus 11.8 (2)
54. standleyanus 13.2 (2)
55. sylvestris 12.4 (4)
56. tamariscinus 12.6 (3)
57. tamaulipensis 13.9 (2)
58. tenuifolius 12.9 (2)
59. thunbergii 14.2 (3)
60. torreyi 13.0 (3)
61. tricolor 13.6 (4)
62. tuberculatus 12.6 (30
63. urceolatus 11.9 (2)
64. venulosus 11.5 (2)
65. viridis 13.9 (5)
66. vulgatissimus 12.2 (3)
67. watsonii 12.3 (3)
68. wrightii 11.9 (2)
7) Alternanthera see Table 2
8) Arthraerua leubnitziae 24.3
9) Blutaparon 13.1
1. portulacoides 13.2 (3)
2. rigidum 11.1 (2)
3. vermiculare 13.2 (3)
4. wrightii 15.0 (2)
10) Bosea 25.4
1. amherstiana 23.3 (2)
2. yervamora 27.4
11) Calicorema 23.3
1. capitata 23.1 (3)
2. squarrosa 23.5 (2)
12) Celosia 28.2
1. anthelminthica 23.9 (3)
2. argentea 28.8 (3)
3. bonnivairii 26.3
Taxon d
13
C%(N, if .1)
4. chenopodifolia NK
5. cristata 28.5
6. elegantissima 29.4
7. fadenorum 25.4
8. floribunda 26.1 (2)
9. globosa 31.8
10. grandifolia 32.6
11. hastata 30.1 (3)
12. isertii 29.7
13. leptostachya 31.1 (2)
14. loandensis NK
15. monosperma 26.8 (2)
16. nervosa NK
17. nitida 28.3
18. orcutti 27.2
19. palmeri 24.2
20. patentiloba 33.2
21. polystachya 25.7
22. pseudovirgata 33.1
23. richardsiae NK
24. schweinfurthiana 28.8 (2)
25. spicata 27.3
26. stuhlmanniana 25.9
27. swinhoei NK
28. trigyna 22.0
29. vanderystii 28.2 (2)
30. virgata 30.4
13) Centrostachys aquatica 26.0 (2)
14) Centema 25.5
1. angolensis 24.6 (2)
2. subfusca 26.4 (2)
15) Centemopsis 26.5
1. fastigiata 27.2 (2)
2. gracilenta 26.9 (2)
3. kirkii 26.4 (3)
4. rubra 25.4 (3)
16) Chamissoa 28.6
1. acuminate 29.2
2. altissima 28.0
17) Charpentiera 27.2
1. australis 22.7
2. densiflora 24.8
3. elliptica 27.8
4. obovata 29.6 (2)
5. ovata 29.2
6. tomentosa 29.2
18) Chionothrix 25.2
1. latifolia 25.0 (3)
2. somalensis 25.4 (2)
19) Cyathula 28.0
1. achyranthoides 31.8 (2)
2. capitata 29.1 (2)
3. cylindrica 27.8 (2)
4. lanceolata 25.1
5. manni NK
6. natalensis 30.0
7. officinalis 27.5
8. orthacantha 26.0
9. polycephala 23.7 (2)
10. prostrata 33.0 (3)
11. tomentosa 26.0 (2)
12. uncinulata 28.2 (2)
20) Dasyspheara 25.0
1. alternifolia 24.4 (2)
2. hyposericea 26.1 (3)
1994 AMERICAN JOURNAL OF BOTANY [Vol. 94
TABLE 1. Continued.
Taxon d
13
C%(N, if .1)
3. robecchii 24.4 (2)
4. tomentosa 25.2 (3)
21) Deeringia 26.0
1. amaranthoides 22.9
2. arborescens 26.0
3. celosiodes 24.3
4. densiflora 27.6 (2)
5. indica 30.2
6. polysperma 24.8 (2)
22) Digera muricata 26.9 (3)
23) Eriostylos stefaninii 25.9 (2)
24) Froelichia 12.0
1. arizonica 12.2 (2)
2. braunii 11.8
3. chacoensis 13.5
4. drummondii 11.5 (3)
5. floridana 12.6 (2)
6. gracilis 12.4 (2)
7. humboldtiana 12.5 (2)
8. interrupta 10.8 (2)
9. juncea 10.1 (2)
10. latifolia 14.2 (2)
11. nudicaulis 10.7 (2)
12. paraguayensis 10.6
13. procera 12.2 (2)
14. texana 13.9 (3)
15. tomentosa 11.6 (2)
16. xantusii 13.5 (2)
25) Froelichiella grisea 25.4 (3)
26) Gomphrena see Table 3
27) Gossypianthus 12.5
1. brittonii 13.4 (2)
2. lanuginosus 11.6 (3)
28) Guilleminia 13.1
1. densa 12.4 (7)
2. chacoensis 13.6
3. elongata 12.2
4. sp. nov. 13.7 (3)
29) Hebanthe 27.0
1. grandiflora 27.5 (3)
2. hookeriana 26.3
3. occidentalis 28.2 (4)
4. paniculata 27.2 (2)
5. reticulata 26.0 (2)
30) Hebanthodes not sampled
31) Henonia scoparia 26.0 (2)
32) Herbstia brasiliana 26.5
33) Hermbstaedtia 25.7
1. angolensi 26.7 (2)
2. argentiformis 26.3 (2)
3. caffra 26.3 (2)
4. fleckii 26.1 (2)
5. glauca 22.8 (2)
6. gregoryi 28.2 (2)
7. linearis 24.0 (2)
8. nigrescens 27.3
9. odorata 25.0 (2)
10. scabra 26.6 (2)
11. schaeferi 23.6
12. spathulaefolia 25.6 (2)
34) Indobanalia thrysiflora 28.2
35) Irenella chrysotricha 25.4 (2)
36) Iresine 25.0
1. alternifolia 22.5 (2)
2. angustifolia 22.0
3. arbuscula 23.9 (2)
4. cassiniiformis 26.3 (2)
5. diffusa 27.9 (4)
6. discolor 23.6 (2)
Taxon d
13
C%(N, if .1)
7. herbstii 26.0 (2)
8. heterophylla 26.6 (2)
9. latifolia 25.0
10. leptoclada 21.3 (2)
11. nigra 30.0 (2)
12. palmeri 28.3 (2)
13. pringlei 23.8 (2)
14. rhizomatosa 26.6 (2)
15. rotundifolia 22.3
16. schaffneri 23.4 (2)
37) Kyphocarpa 24.9
1. angustifolia 25.9 (2)
2. trichinoides 23.9
38) Lagrezia 30.2
1. boivinii 28.8
2. madagascariensis 29.8 (3)
3. micrantha 33.7
4. oligomeroides 28.5 (3)
39) Lecosia formicarum 36.3
40) Leucosphaera bainesii 24.8 (3)
41) Lithophila 12.1
1. muscoides 12.2 (2)
2. radicata 11.4 (2)
3. subscapulosa 12.7
42) Lopriorea ruspolii 27.1 (2)
43) Marcelliopsis 23.7
1. denudata 23.9 (3)
2. dinteri 23.2
3. splendens 23.1 (2)
4. welwitschii 24.7 (3)
44) Mechowia 25.4
1. grandiflora 24.2 (2)
2. redactifolia 26.4
45) Nelsia quadrangula 25.3 (3)
46) Neocentema 26.2
1. alternifolia 26.4 (2)
2. robecchii 26.0 (2)
47) Nothosaerva brachiata 29.3
48) Nototrichum 26.2
1. humile 28.9
2. sandwicense 23.3
3. viride 26.5
49) Nyssanthes 27.9
1. diffusa 29.4
2. erecta 26.4
50) Omegandra kanisii 24.2
51) Pandiaka 24.7
1. angustifolia 26.0 (2)
2. carsonii 25.9
3. confusa NK
4. elegantissima 26.0
5. involucrata 26.7 (3)
6. lanuginosa NK
7. porphyrargyrea NK
8. ramulosa 25.6 (2)
9. rubro-lutea 28.4
10. trichinioides 25.4 (2)
11. welwitschii 27.5
52) Pedersenia 25.3
1. argentata 26.1 (2)
2. aurata 25.5
3. cardenasii 25.1
4. hassleriana 24.3 (2)
5. sp. (ined.) sub
Pfaffia completa
NK
6. sp. nova 27.6
7. weberbaueri 22.8
53) Pfaffia 26.8
1. acutifolia 29.0
Taxon d
13
C%(N, if .1)
2. elata 25.5
3. fruticulosa 30.6
4. glabrata 24.2
5. glabratoides 29.5
6. glauca 26.4
7. glomerata 27.9
8. gnaphaloides 25.7
9. helichrysoides 24.8
10. iresinoides 28.9 (2)
11. jubata 26.6 (2)
12. luzulaeflora 23.3 (3)
13. sericea 27.4
14. stenophylla 31.7
15. tenuis NK
16. townsendii 24.3 (2)
54) Pleuropetalum 29.4
1. darwinii 26.3 (2)
2. pleiogynum 32.4
3. sprucei 29.6
55) Pleuropterantha 26.6
1. revoilii 24.8 (3)
2. thulinii 27.4 (2)
3. undulatifolia 27.6
56) Polyrhabda atriplicifolia 25.3
57) Pseudogomphrena scandens 24.6
58) Pseudoplantago 29.8
1. bisteriliflora 29.3
2. friesii 30.3
59) Pseudosericocoma not sampled
60) Psilotrichopsis 29.7
1. cochinchinensis 29.7 (3)
2. curtissii NK
61) Psilotrichum 27.1
1. africanum 26.7 (2)
2. amplum 26.0 (2)
3. axilliflorum 31.4 (2)
4. cordatum 24.7 (2)
5. cyathuloides 28.5
6. elliottii 27.2 (2)
7. erythrostachyum 26.9 (2)
8. ferrugineum 26.0 (2)
9. gloveri 27.3 (2)
10. gnaphalobryum 29.7 (2)
11. gracilipes 24.8
12. lanatum 25.6 (2)
13. leptostachyum 26.9
14. nudum 24.2 (2)
15. pedunculosum 27.1 (2)
16. schimperi 27.8 (2)
17. scleranthum 29.7 (2)
18. stenanthum 26.4 (2)
19. tomentosum 27.1 (2)
20. virgatum 27.2 (4)
62) Ptilotus 25.3
1. alopecuroideus 24.4 (2)
2. astrolasius 22.5 (2)
3. auriculifolius 23.8
4. calostachyus 23.7 (2)
5. conicus 28.0 (2)
6. corymbosus 24.7 (2)
7. declinatus 23.3
8. dissitiflorus 26.0 (2)
9. distans 25.8 (2)
10. divericatus 24.2 (2)
11. drummondii 25.0 (2)
12. erubescens 25.2 (2)
13. esquamatus 25.2 (2)
14. fusiformis 27.4 (2)
15. gaudichaudii 26.4 (2)
December 2007] SAGEETAL.—C
4
PHOTOSYNTHESIS IN THE AMARANTHACEAE 1995
being C
4
species (Tables 1–3). Ten of the 77 genera listed in
Table 1 contained C
4
species. Three of these 10 genera (Aerva,
Alternanthera, and Gomphrena) contained both C
3
and C
4
species, while the rest (Amaranthus, Blutaparon,Froelichia,
Gossypianthus,Guilleminea,Lithophila, and Tidestromia) had
only C
4
species. Aerva had four C
4
species and 10 C
3
species
(Table 1). Alternanthera had 15 C
4
species, representing 17%
of the sampled taxa in the genus (Table 2). Gomphrena had
109 C
4
species of 122 examined, representing 89% of the
sampled taxa in the genus (Table 3). The 13 C
3
Gomphrena
species were all from the Americas. All native Australian
Gomphrena species were C
4
(Table 3).
The isotope values of the sampled Amaranthaceae species
segregated into two distinct distributions (Fig. 1). None of the
607 species had isotope ratios between 17%and 21%,
which might indicate the presence of C
4
-like C
3
–C
4
interme-
diate species (Sage, 2004). The frequency distribution of
isotope values of C
3
and C
4
species in Aerva and Gomphrena
was similar to the frequency distribution of all C
3
and C
4
species in the study, indicating no trend toward an intermediate
C
3
–C
4
physiology (compare Figs. 1 and 2). By contrast, five
species in Alternanthera had isotope values between 21.5 and
23.1%, a range that corresponded to the upper fringe of all
the C
3
species in the study. An isotope value above 24
indicates either high water use efficiency or the potential for
PEP carboxylase engagement in C
3
–C
4
intermediacy (Farquhar
et al., 1989; Monson and Rawsthorne, 2000). All three
Alternanthera species previously shown to be C
3
–C
4
interme-
diates (Al. crucis, Al. ficoidea,Al. tenella; Rajendrudu et al.,
1986; Rajendrudu and Das, 1990; Fernandez et al., 1999) had
C
3
-like carbon isotopic values that averaged 28%(Table 2).
A number of species deserve note. Three C
4
Amaranthaceae
species occur at high elevations (.4000 m a.s.l.) where C
4
plants are rarely observed. Gomphrena meyeniana (13.7%)
occurs in the Argentinian and Bolivian Andes at 3800 to 4600
m a.s.l. (Pedersen, 1990; Borsch et al., in press). The 4600 m
collection of G. meyeniana examined here (Solomon, Stein and
Uehlig 11794 from Valle del Zongo in Bolivia) is the highest-
elevation C
4
eudicot known in the world. The altitude record
reported for any confirmed C
4
species worldwide is 4800 m for
the grass species Muhlenbergia peruviana (Ruthsatz and
Hoffmann, 1984). This species and G. meyeniana are the only
known C
4
plants above 4500 m in the western hemisphere. In
the eastern hemisphere, the chenopod Salsola monoptera and
the grass Orinus thoroldii are the only putative C
4
species
reported above 4500 m; both occur on the Tibetan Plateau
(Wang, 2003). Gomphrena umbellata (12.9%) and Am.
peruvianus (13.8%) occur from 3700 to 4300 m in the central
Andes of Bolivia and Argentina (Pedersen, 1990; Borsch et al.,
in press). Gomphrena meyeniana and Al. peruvianus have a
typical alpine plant morphology, with small leaves tightly
clustered on a thick root stock, whereas G. umbellata is an
annual growing in dry sand fields. A fourth high-elevation C
4
species that we noted is Al. microphylla (12.4%), which
grows up to 4000 m a.s.l. in dry chaparral vegetation of the
central Andes (Beck et al., 2001; Borsch et al., in press). A C
3
species of Alternanthera also occurs in the Peruvian Andes
above 4000 m a.s.l. (Al. lupulina, 27.4%). These two
Alternanthera species have a similar growth form and would
be ideal for comparing C
3
and C
4
photosynthetic performance
at high elevation.
Parsimony analysis of combined matK/trnK sequences
yielded 809 shortest trees of 2617 steps (CI ¼0.620, RC ¼
0.466). The majority rule strict consensus is shown in Fig. 3,
and one of the shortest trees is in Fig. 4. In Amaranthaceae, the
basal grade consists of Bosea, followed by Charpentiera. The
additional major lineages correspond to the tribe Celosieae, the
Amaranthoid clade (Amaranthus and relatives), Psilotrichum
ferrugineum and Allmaniopsis in isolated positions, the
Aervoid clade (Aerva and relatives), and the Achyranthoid
clade (containing mostly paleotropical genera). The largely
neotropical subfamily Gomphrenoideae appears with the
lineages of Iresine,Alternanthera,Tidestromia,Pedersenia,
and Pseudoplantago in a polytomy with a well-supported clade
comprising Blutaparon (C
4
), Froelichia (C
4
), Gomphrena (C
4
),
Guilleminea (C
4
), Hebanthe (C
3
), Pfaffia (C
3
), and the C
3
genus Xerosiphon (Fig. 3).
TABLE 1. Continued.
Taxon d
13
C%(N, if .1)
16. grandiflorus 27.0 (2)
17. helipteroides 22.8 (2)
18. holosericeus 25.3 (2)
19. humulis 28.4 (2)
20. incanus 24.3 (2)
21. lanatus 22.2 (2)
22. latifolius 22.7 (2)
23. leucocomus 27.7 (2)
24. macrocephalus 27.5 (2)
25. manglesii 26.0 (2)
26. nobilis 26.1 (2)
27. obovatus 22.1 (2)
28. polystachyus 26.4 (2)
29. pyramidatus 23.8
30. reductifolia 26.4
31. rotundifolius 25.5 (2)
32. schwartzii 26.1 (2)
33. sericostachyus 26.8 (2)
34. spathulatus 24.9 (2)
35. spicatus 25.9 (2)
Taxon d
13
C%(N, if .1)
36. stirlingii 26.3 (2)
37. villosiflorus 25.0 (2)
63) Pupalia 25.4
1. lappacea 25.1 (6)
2. mollis 25.8
3. orbiculata 27.3
4. schimperiana 23.5
5. sericea 25.2
64) Quaternella ephedroides 25.5
65) Rosifax sabuletorum 28.1
66) Saltia papposa 26.2 (3)
67) Sericocoma 24.0
1. avolans 24.9
2. heterochiton 23.1
68) Sericocomopsis 23.1
1. hildebrandtii 23.0 (2)
2. pallida 23.2 (2)
69) Sericorema 24.5
1. remotiflora 23.5
2. sericea 25.5
Taxon d
13
C%(N, if .1)
70) Sericostachys scandens 31.4
71) Siamosia thailandica 24.1
72) Stilbanthus scandens 25.6 (2)
73) Tidestromia 13.0
1. carnosa 12.4 (3)
2. gemmata 12.7 (3)
3. lanuginosa 13.0 (3)
4. oblongifolia 13.5 (3)
5. rhizomatosa 12.5
6. suffruticosa 12.8 (3)
7. tenella 12.7 (2)
8. valdesii 14.5
74) Trichuriella monsoniae 27.2 (4)
75) Volkensinia prostrata 24.8 (2)
76) Woehleria serpyllifolia 34.2 (2)
77) Xerosiphon 25.8
1. angustiflorus 25.8 (2)
2. aphyllus 25.7 (5)
1996 AMERICAN JOURNAL OF BOTANY [Vol. 94
DISCUSSION
Recent molecular studies have provided novel insights into
phylogenetic relationships within the Amaranthaceae (Kadereit
et al., 2003; Pratt, 2003; Mu¨ ller and Borsch, 2005a, b). Here,
we analyzed the matK/trnK sequences from Mu¨ ller and Borsch
(2005a, b), matK/trnK data from Iresine (T. Borsch et al.,
unpublished data) and several new matK/trnK sequences from
Aerva, Alternanthera, and Amaranthus species to generate a
more complete picture of relationships between C
3
and C
4
taxa
within the Amaranthaceae (Figs. 3, 4). The topology of our
matK/trnK tree is congruent with the trees found by Mu¨ller and
Borsch (2005a, b); however, some nodes that were weakly
supported in their analyses (e.g., Deeringia as sister to the
remainder of Celosioid genera, and some deep nodes in
Gomphrenoideae) were not resolved with statistical support
when more taxa were added in this study (shown as polytomies
in the majority rule consensus of Fig. 3).
Five different lineages with C
4
photosynthesis are indicated.
The first is the C
4
-only genus Amaranthus, which is depicted to
be monophyletic within a monophyletic Amaranthoid lineage
(Figs. 3, 4). Our addition of the genus Pleuropterantha shows
that this drought-adapted tropical herb from Africa is sister to
Amaranthus rather than to Chamissoa (a genus of woody
shrubs and lianas of wet forests in the neotropics) as found in
the previous phylogenetic analyses by Mu¨ ller and Borsch
(2005a, b). This finding is consistent with hypotheses that C
4
photosynthesis evolved in dry regions (Sage, 2004) and
suggests that closer phylogenetic and ecophysiological exam-
ination of Amaranthus,Pleuropterantha, and putative relatives
such as Digeria may provide clues to the origin of C
4
photosynthesis in this clade.
The second C
4
group is found in the aervoids and is
constituted by a lineage of Aerva javanica and closely related
species that appear sister to the herbaceous genus Nothosaerva.
TABLE 2. Photosynthetic pathways in Alternanthera. Boldface indicates
C
4
taxa. INT indicates species previously identified as C
3
C
4
intermediate species. Data are d
13
C values (followed by sample size
where N .1). See Appendix S2 (with Supplemental Data in the
online version of this article) for samples and collection information.
Seventeen (19.5%) of the 87 accepted taxa listed below are C
4
. The
mean d
13
C value for C
3
Alternanthera species is 27.2%and for C
4
Alternanthera species is 12.2%. The mean d
13
C value for the three
C
3
–C
4
species is 28.0%.
Alternanthera species d
13
C%(N, if .1)
1. albida 14.2 (3)
2. albosquarrosa 25.7 (2)
3. albotomentosa 27.2 (3)
4. altacruzensis 26.4 (3)
5. amoena 26.7 (2)
6. aquatica 27.8 (2)
7. areschougii 30.7 (2)
8. bettzickiana 28.8 (2)
9. boliviana 12.8
10. brasiliana 28.0 (4)
11. canescens 12.6 (3)
12. caracasana 14.4 (2)
13. chacoensis 11.2
14. cinerella 10.8 (2)
15. congesta 27.0
16. costaricensis 33.6 (2)
17. crucis INT 28.0 (3)
18. dolichocephala 23.1
19. dominii 28.7 (2)
20. echinocephala 22.9 (4)
21. eggersii 26.4 (3)
22. elongata 27.1 (5)
23. eupatoroides 22.2 (3)
24. ficoidea INT 27.7 (4)
25. filifolia 25.0 (5)
26. flava 28.2 (2)
27. flavescens 27.8 (5)
28. flavicoma 29.4 (2)
29. flavida 27.1
30. flavogrisea 28.8 (3)
31. frutescens NK
32. genticulata 30.3
33. gracilis 26.5 (2)
34. halimifolia 26.3 (2)
35. helleri 24.9 (4)
36. hirtula 29.2 (2)
37. jacquinii 28.9
38. juniciflora 28.6
39. kuntzii 28.4 (2)
40. kurtzii 28.2 (3)
41. laguroides 27.0 (4)
42. lanceolata 30.2 (2)
43. lehmannii 29.5 (2)
44. lupulina 26.7 (2)
45. macbridei 25.5 (5)
46. malmeana 27.0 (3)
47. maritima 26.3 (4)
48. markgrafii 26.7
49. martii 27.5 (4)
50. micrantha 28.0 (3)
51. microcephala 28.3
52. microphylla 12.4 (4)
53. moquinii 25.4 (3)
54. multicaulis 10.2
55. nesiotes 10.4
56. nifa 25.6
57. nodifera 13.6
58. obovata 28.8
59. olivacea 26.8
60. paniculata 24.2 (3)
61. paronychioides 11.5 (4)
TABLE 2. Continued.
Alternanthera species d
13
C%(N, if .1)
62. peruviana 12.1 (3)
63. philoxeroides 29.6 (2)
64. porrigens 26.2 (5)
65. portoricensis 28.2 (3)
66. publiflora 27.3 (2)
67. pulchella 11.3 (3)
68. pumilla 13.0 (2)
69. pungens 12.3 (2)
70. pycnantha 27.4 (2)
71. reineckii 28.4 (2)
72. repens 11.4
73. rufa 25.8 (2)
74. rugulosa 27.7
75. scandens 29.0 (4)
76. serpens 13.9
77. serpyllifolia 28.3
78. sessilis 28.5 (3)
79. snodgrassii 21.6
80. spinosa 22.7 (2)
81. stellata 27.8 (8)
82. tenella INT 28.4 (6)
83. truxillensis 26.6 (5)
84. tucumana 29.5 (2)
85. vaga 24.9
86. vestita 25.0 (2)
87. villosa 26.9 (3)
December 2007] SAGEETAL.—C
4
PHOTOSYNTHESIS IN THE AMARANTHACEAE 1997
TABLE 3. Photosynthetic pathways in Gomphrena. Boldface indicates C
4
taxa. Data are d
13
C values (followed by sample size where N .1). K
indicates Kranz anatomy was observed. See Appendix S3 (with
Supplemental Data in the online version of this article) for complete
list of samples and collection information. 109 (89%) of the 122 listed
species are C
4
. Mean d
13
C values are 27.3 for the C
3
Gomphrena
and 13.3 for the C
4
Gomphrena. The mean d
13
C value for New
World C
4
Gomphrena is 13.1; Australian mean d
13
C value is 13.7
(these differences were significant at P,0.05).
Gomphrena species d
13
C%(N, if .1)
New World Species (86%C
4
)
1. aborescens 13.1 (2)
2. agrestis 13.8 (6)
3. albiflora 13.4
4. aurea 10.8 (2)
5. basilanata 12.8
6. bicolor 16.2
7. blanchetii 13.3
8. boliviana 11.7 (2)
9. caespitosa 12.1 (2)
10. celosiodes 14.3 (2)
11. centrota 13.8 (2)
12. cladotrichoides 12.2
13. claussenii 25.8
14. debilis 14.8
15. decipiens 12.4
16. decumbens 13.5 (3)
17. demissa 14.0
18. desertorum 13.4 (2)
19. discolor 11.6
20. dispersa 10.7
21. duriuscula 13.9 (2)
22. elegans 27.3 (3)
23. ferruginea 15.2 (2)
24. filaginoides 11.2
25. fuscipellitea 14.2
26. gardneri 13.5
27. globosa 13.1 (2)
28. gnaphiotricha 28.1
29. graminea 13.5 (2)
30. guaranitica 11.8 (2)
31. haageana 12.7
32. haenkeana 12.1 (3)
33. hassleri 12.2
34. hermogenesii 25.8
35. hillii 13.2 (2)
36. holosericea 29.3 (2)
37. hygrophila 12.4 (2)
38. incana 12.8 (2)
39. lanigera 12.5
40. lutea 10.6
41. leucocephala 13.3
42. macrocephala 12.3 (3)
43. mandonii 25.5 (3)
44. martiana 12.6 (3)
45 mendocina 11.9 (2)
46. meyeniana 13.7 (3)
47. microcephala K
48. mollis 28.0 (2)
49. moquinii 12.6 (2)
50. nana 13.1 (2)
51. nealleyi 12.7 (2)
52. nitida 12.8 (2)
53. officinalis 12.5 (2)
54. oligocephala 11.1 (2)
55. oroyana 13.6 (2)
56. pallida 14.3
57. paraguayensis 27.8 (2)
58. paranensis 14.0 (3)
59. perennis 14.7 (3)
TABLE 3. Continued.
Gomphrena species d
13
C%(N, if .1)
60. phaeotricha 12.4
61. pilosa 28.8 (2)
62. pohlii 14.7 (2)
63. pringlei 12.3 (2)
64. procumbens 13.3 (2)
65. prostrata 14.8 (2)
66. pulchella 13.0 (2)
67. pulcherrima 13.8 (2)
68. pulvinata 14.5
69. pumila 12.0
70. pungens 14.9
71. radiata 13.2 (2)
72. regeliana 12.3 (2)
73. riedelliana 11.4
74. rhodantha 10.9
75. rudis 13.4 (2)
76. rupestris 27.1 (4)
77. scapigera 12.8 (2)
78. schlechtendaliana 12.8
79. serrata 13.8 (2)
80. serturneroides 11.5
81. silenoides 12.2 (2)
82. sonorae 14.1
83. spissa 11.8
84. stellata 15.1
85. tomentosa 11.0 (2)
86. trollii 14.1
87. tuerckheimii 26.1 (2)
88. umbellata 12.9 (3)
89. unitgincensis 13.3
90. vaga 26.2 (3)
91. virgata 12.6 (2)
92. viridifolia 27.6
Native Australian Species (100%C
4
)
93. affinis 12.8 (3)
94. arida 11.3
95. atrorubra 13.8 (3)
96. brachystylis 13.5 (4)
97. brevifolia 13.6 (4)
98. canescens 13.4 (4)
99. conferta 13.8
100. conica 15.5 (3)
101. connata 14.1 (3)
102. cunninghamii 13.1 (2)
103. diffusa 14.2 (3)
104. eichleri 14.2 (2)
105. flaccida 14.2 (3)
106. floribunda 13.7 (3)
107. humifusa 14.6
108. humilis 14.1 (3)
109. involucrata 14.4 (3)
110. kanisii 13.1
111. lacinulata 13.3 (3)
112. lanata 13.2 (4)
113. leontopodoides 13.7
114. leptoclada 12.5 (4)
115. leptophylla 14.6 (4)
116. magentipetala 14.1 (4)
117. occulta 14.3
118. parviflora 14.4 (4)
119. pusilla 14.5
120. rosula 14.6 (4)
121. sordida 12.2
122. tenella 13.9
1998 AMERICAN JOURNAL OF BOTANY [Vol. 94
This study shows that most C
3
species of Aerva occur in a
different clade that is sister to Ptilotus.Aerva is therefore
determined to be polyphyletic. This possibility was indicated
by the respective positions of Ae. javanica and Ae. leucura in
the analyses of Mu¨ller and Borsch (2005a, b). In a recent
analysis of Aerva, Thiv et al. (2006) also found two major
clades in the genus. However, because Aerva was rooted with
Ptilotus in their study, its polyphyly was not evident. One
branch of Aerva, termed clade B by Thiv et al. (2006), is
completely C
3
, with species Ae. coriacea,Ae. congesta, Ae.
lanata,Ae. leucura,Ae. sanguinolenta, and Ae. triangular-
ifolia. The C
4
species of Aerva are segregated into a separate
clade, along with the xerophytic C
3
species Ae. artemisoides,
Ae. microphylla, and Ae.revoluta (Thiv et al., 2006). These C
3
species should be close to the ancestor of the C
4
species and
thus may reveal insights into the evolution of C
4
photosynthe-
sis in Aerva.
All remaining C
4
groups of Amaranthaceae are in the largely
new world subfamily Gomphrenoideae, which is monophyletic
(Mu¨ller and Borsch, 2005a; I. Sa´nchez-del Pino, unpublished
data). This third and most speciose C
4
clade of Amaranthaceae
comprises the genera Froelichia,Guilleminea,Blutaparon,C
4
Gomphrena (Fig. 3), and most likely also Gossypianthus and
Lithophila. The latter two are not included in the phylogenetic
analysis in this study, but pollen and morphological evidence
strongly indicates that these two groups are members of this C
4
clade (Borsch, 1998). Lithophila also shares many morpho-
logical characters with Blutaparon (Eliasson, 1988). Gom-
phrena has been suggested to be polyphyletic based on pollen
(Borsch, 1998) and sequence data (Kadereit et al., 2003; Mu¨ ller
and Borsch, 2005a). Notably, the C
3
species of Gomphrena are
unrelated to the core of the genus Gomphrena, which appears
to be strictly C
4
(Fig. 3). Our data support the view of Pedersen
(1990) that the C
3
species formerly classified as G. angustifolia
and G. aphyllus should be classified as members of the genus
Xerosiphon. Species of Xerosiphon have metareticulate pollen
with reduced tecta similar to C
4
Gomphrena species. By
contrast, the C
3
species classified as G. mandonii and G.
elegans have metareticulate pollen with complete tecta as
occurs in Pfaffia (Borsch, 1998; T. Borsch lab, unpublished
data). In the molecular tree, the respective C
3
Gomphrena
species also appear as close relatives to Pfaffia and should
eventually be reclassified within Pfaffia or a closely related
genus.
The C
3
genus Froelichiella has yet to be included in any
molecular phylogenetic analysis. Eliasson (1988) has pointed
out that among Gomphrenoideae, Froelichiella is closest to
Froelichia, indicating that it may be the nearest living C
3
relative of the C
4
lineage. There are notable differences
between Froelichiella and Froelichia, however. For example,
in Froelichiella the tepals are almost free to the base, the
stigma is multilobate, and pseudostaminodia are distinct.
Pollen morphology of Froelichiella is of the Gomphrena type
(Borsch, 1998) but has details very close to Xerosiphon (T.
Borsch lab, unpublished data). A phylogenetic position for
Froelichiella that is either sister to or diverging near
Xerosiphon is thus hypothesized.
The fourth postulated C
4
lineage is constituted by the
monophyletic genus Tidestromia, which has a center of
diversity in gypsum soils in Mexico and the United States
(Sa´nchez-del Pino and Flores Olvera, 2006). This is a small
genus with eight C
4
species. Current phylogenetic hypotheses
raise the possibility of either Alternanthera and Tidestromia
being sister groups (rbcL sequence data, Kadereit et al., 2003;
trnL-F sequence data, I. Sa´nchez-del Pino, unpublished data)
or Pedersenia and Alternanthera being sister groups (rpl16
sequence data, rpl16þtrnLF sequence data combined, I.
Sa´nchez-del Pino, unpublished data). However, statistical
support for either hypothesis is weak, and matK/trnK data of
this study (Fig. 3) depict Alternanthera,Pedersenia, and
Tidestromia in a polytomy. The addition of Al. altacruzensis,
Al. flavescens, and Al. microphylla to the matK/trnK data set in
this study indicates that C
4
photosynthesis in Alternanthera
arose in a terminal clade of the genus, well after the
diversification of major Alternanthera lineages. Alternanthera
therefore is the fifth independent C
4
lineage in Amaranthaceae.
Recent work using trnLF and rpl16 sequence data and more
Fig. 1. The frequency distribution of carbon isotope ratios for the
species listed in Tables 1–3. Carbon isotope ratio was examined on 596 of
the 607 species in the study. Fig. 2. The frequency distribution of carbon isotope ratios for the C
3
and C
4
species in the genera Alternanthera, Gomphrena and Aerva listed
in Tables 1–3.
December 2007] SAGEETAL.—C
4
PHOTOSYNTHESIS IN THE AMARANTHACEAE 1999
2000 AMERICAN JOURNAL OF BOTANY [Vol. 94
species indicates Alternanthera to be monophyletic (I.
Sa´nchez-del Pino, unpublished data) and also suggests C
3
photosynthesis to be the plesiomorphic condition in the genus.
The C
4
and C
3
–C
4
intermediate species of Alternanthera
appear to occur in one subclade of procumbent herbs that form
a terminal clade relative to the C
3
Alternanthera species. These
species are largely South American, in contrast to Tidestromia,
which is centered in northern Mexico and the southwestern
USA. This difference in the center of diversity between the C
4
Alternanthera species and Tidestromia also supports the
hypothesis that C
4
photosynthesis independently arose in these
two genera. A distinct origin of C
4
photosynthesis in
Alternanthera is also supported by the occurrence of three
species in the genus with biochemical and anatomical traits that
are intermediate between the full C
3
and C
4
conditions
(Fernandez et al., 1999). The exact phylogenetic position of
these intermediates, however, is not currently known. Because
Alternanthera is species rich compared to other genera with
identified C
3
–C
4
species, it may be an excellent system to
study the evolution of the C
4
pathway. A priority for future
work with Alternanthera should be clarification of the
phylogenetic positions of the C
3
,C
3
–C
4
, and C
4
species.
C
3
–C
4
intermediate species are classified into two function-
ally distinct groups. The first group comprises intermediate
species that have an efficient system to recapture photorespired
CO
2
. This occurs by localizing the photorespiratory enzyme
glycine decarboxylase in thebundlesheathtissueand
transporting all photorespiratory metabolites formed in a leaf
to this compartment for decarboxylation (Monson and Raws-
thorne, 2000; Sage, 2004). The released CO
2
accumulates in
the bundle sheath tissue and is refixed by Rubisco present in
these cells. These species are termed type I C
3
–C
4
intermedi-
ates (Edwards and Ku, 1987). Type I intermediates have a
typical C
3
isotopic signature because all CO
2
fixation occurs
via Rubisco and the d
13
C ratio reflects isotopic discrimination
by Rubisco. Thus, isotopic screens cannot identify type I
intermediates, which include the three Alternanthera interme-
diates previously identified by biochemical and anatomical
studies (Rajendrudu et al., 1986; Rajendrudu and Das, 1990;
Devi and Raghavendra, 1993). The second group (type II
intermediates sensu Edwards and Ku, 1987) has enhanced
activity of PEP carboxylase and a limited C
4
cycle. In type II
intermediates, initial fixation of CO
2
by PEP carboxylase
increases the d
13
C ratio toward C
4
-like values; however, to rise
above the 21%threshold that excludes C
3
species, the
activity of PEP carboxylase has to be substantially enhanced,
typically above 50% of the activity of Rubisco (Monson et al.,
1988; Monson and Rawsthorne, 2000). We did not identify any
species with d
13
C ratios between 17%to 21%that are
strong indicators of type II intermediacy; however, we did
identify a number of Alternanthera species (Al. dolichocepha-
la, Al. echinocephala, Al.eupatoroides, Al. snodgrassii, and Al.
spinosa) with d
13
C values that stood out because they were on
the upper fringe of the C
3
distribution. These species would be
excellent candidates to examine for type II intermediacy.
Alternatively, these species may have low stomatal conduc-
tances often seen in xeric species; low stomatal conductance
relative to photosynthetic capacity generally explains high
isotope ratios in C
3
species (Farquhar et al., 1989).
This survey identified 233 C
4
species of the 607 examined in
the Amaranthaceae. With perhaps two dozen C
4
species not
examined in the survey, we estimate there are about 250 C
4
species in the Amaranthaceae or about 28% of the 900 species
estimated to occur in the family (Townsend, 1993). This value
is identical to the 250 C
4
species previously estimated by Sage
et al. (1999). Sage et al. (1999) listed Achyranthes and Celosia
as having C
4
species, based on earlier reports of Kranz
anatomy. Our results show no evidence of C
4
photosynthesis in
either Achyranthes or Celosia, although enlarged bundle sheath
cells may indicate C
3
–C
4
intermediacy.
In conclusion, we provide the first comprehensive survey of
the distribution of the photosynthetic pathways within a large
Fig. 3. Majority rule parsimony tree based on matK/trnK sequence data depicting relationships in Amaranthaceae with clade annotations and isotope
values. Isotope values for the species are indicated in brackets, if available. Isotope values that do not correspond to values in Tables 1–3 are from the
specimen sampled for DNA.
Fig. 4. One of the shortest phylogenetic trees showing the distribution
of the C
4
pathway in the Amaranthaceae s.s. The C
4
branches are indicated
in bold.
December 2007] SAGEETAL.—C
4
PHOTOSYNTHESIS IN THE AMARANTHACEAE 2001
eudicot family that evolved C
4
photosynthesis multiple times.
By clarifying which species are C
3
, in addition to which are C
4
,
this survey provides important information for a range of
disciplines for which the photosynthetic pathway has important
consequences. Evolutionary studies will benefit by having the
proper placement of the C
4
pathway in a phylogenetic
sequence. With the development of fine scale phylogenies,
our results can be used to identify closely related C
3
and C
4
species, which would then facilitate studies on the evolutionary
changes involved in C
4
plant evolution. By knowing with
certainty whether a species is C
3
or C
4
, ecologists will be able
to identify the ecological significance of a photosynthetic
pathway in a given flora and within a vegetation community. In
combination with evolutionary studies, ecological research
with closely related C
3
and C
4
species could identify in greater
detail the ecological conditions favoring the rise of C
4
photosynthesis. The current understanding is that C
4
species
arose in hot, arid or saline environments, but no study has
clearly documented the local ecological conditions where any
specific C
4
lineage first appeared; hence, the generalization
remains untested. Our survey has also identified species of
interest to physiological ecologists. For example, the identifi-
cation of C
3
and C
4
Alternanthera species from high elevations
and of Gomphrena meyeniana as the world’s highest-elevation
C
4
eudicot provides a robust system for examining adaptation
of the C
4
pathway for cold climates that are markedly different
from the hot environments where the pathway is believed to
have first evolved.
LITERATURE CITED
AKHANI, H., P. TRIMBORN,AND H. ZIEGLER. 1997. Photosynthetic pathways
in Chenopodiaceae from Africa,AsiaandEuropewiththeir
ecological, phytogeographical and taxonomical importance. Plant
Systematics and Evolution 206: 187–221.
ANGIOSPERM PHYLOGENY GROUP (APG). 2003. An update of the angiosperm
phylogeny group classification for the orders and families of
flowering plants: APG II. Botanical Journal of the Linnean Society
141: 399–436.
BAILLON, H. E. 1887. Histoire des plantes. Hachette, Paris, France.
BECK, S. N., Z. M. PANIAGUA,G.YEVARA,AND M. LIBERMAN. 2001. La
vegetacio´ n y uso de la tierra del altipiano y de los valles en el oeste
del departamento de Tarija, Bolivia. In S. Beck, N. Paniagua, and D.
Preston [eds.], Historia, Ambiente y Sociedad en Tarija, Bolivia, 47–
94. Editorial Instituto de Ecologı´a, La Paz, Bolivia.
BORSCH, T. 1993. Amaranthaceae. In L. Brako and J. L. Zarucchi [eds.],
Catalogue of the flowering plants and gymnosperms of Peru, 18–26.
Missouri Botanical Garden Press, St. Louis, Missouri, USA.
BORSCH, T. 1998. Pollen types in the Amaranthaceae—morphology and
evolutionary significance. Grana 37: 129–142.
BORSCH, T., T. ORTUN
˜O,AND M. NEE. In press. Amaranthaceae. In P. M.
Jorgensen and S. G. Beck [eds.], Cata´logo de las plantas vasculares de
Bolivia. Monographs in systematic botany. Missouri Botanical
Garden, St. Louis, Missouri, USA.
BORSCH, T., AND T. M. PEDERSEN. 1997. Restoring the generic rank of
Hebanthe Martius (Amaranthaceae). Sendtnera 4: 13–31.
BORSCH, T., AND D. QUANDT. 2007. Mutational dynamics and phylogenetic
utility of chloroplast non-coding DNA. Plant Systematics and
Evolution Special Volume on molecular evolution and phylogenetic
utility of non-coding DNA. In press.
CAVACO, A. 1954. Amaranthaceae. In H. Humbert [ed.], Flore de
Madagascar et des Comores, 1–56. Firmin-Didot, Paris, France.
CUE
´NOUD, P., V. SAVOLAINEN,L.CHATROU,M.POWELL,R.J.GRAYER,AND
M. W. CHASE. 2002. Molecular phylogenetics of Caryophyllales
based on nuclear 18S rDNA and plastid rbcL,atpB, and matK DNA
sequences. American Journal of Botany 89: 253–273.
DEVI, M. T., AND A. S. RAGHAVENDRA. 1993. Partial reduction in activities
of photorespiratory enzymes in C
3
–C
4
intermediates of Alternanthera
and Parthenium.Journal of Experimental Botany 44: 779–784.
DOWNIE, S. R., D. S. KATZ-DOWNIE,AND K. Y CHO. 1997. Relationships in
the Caryophyllales as suggested by phylogenetic analyses of partial
chloroplast DNA ORF2280 homolog sequences. American Journal of
Botany 84: 253–273.
DOWNTON, W. J. S. 1975. Occurrence of C
4
photosynthesis among plants.
Photosynthetica 9: 96–105.
EDWARDS, G. E., AND M. S. B. KU. 1987. Biochemistry of C
3
–C
4
intermediates. In P.K.StumpfandE.E.Conn[eds],The
biochemistry of plants, 275–325. Academic Press, San Diego,
California, USA.
EHLERINGER, J. R., T. E. CERLING,AND B. R. HELLIKER. 1997. C
4
photosynthesis, atmospheric CO
2
and climate. Oecologia 112: 285–
299.
ELIASSON, U. H. 1987. Amaranthaceae. In G. Harling and L. Andersson
[eds.], Flora of Ecuador, no. 28, Berlings, Arlo¨v, Gothenburg,
Sweden.
ELIASSON, U. H. 1988. Floral morphology and taxonomic relations among
the genera of Amaranthaceae in the New World and the Hawaiian
Islands. Botanical Journal of the Linnean Society 96: 235–283.
FARQUHAR, G. D., J. R. EHLERINGER,AND K. T. HUBICK. 1989. Carbon
isotope discrimination and photosynthesis. Annual Review of Plant
Physiology and Plant Molecular Biology 40: 503–537.
FERNANDEZ, M. D., A. PIETERS,M.AZKUE,E.RENGIFO,W.TEZARA,F.I.
WOODWARD,AND A. HERRERA. 1999. Photosynthesis in plants of four
tropical species growing under elevated CO
2
.Photosynthetica 37:
587–599.
KADEREIT, G., T. BORSCH,K.WEISING,AND H. FREITAG. 2003. Phylogeny of
Amaranthaceae and Chenopodiaceae and the evolution of C
4
photosynthesis. International Journal of Plant Sciences 164: 959–
986.
KAPRALOV, M. V., H. AKHANI,E.V.VOZNESENSKAYA,G.EDWARDS,V.
FRANCESCHI,AND E. H. ROALSON. 2006. Phylogenetic relationships in
the Salicornioideae/Suaedoideae/Salsoloideae s.l. (Chenopodiaceae)
clade and a clarification of the phylogenetic position of Bienertia and
Alexandra using multiple DNA sequence data sets. Systematic Botany
31: 571–585.
KU
¨HN, U., V. BITTRICH,R.CAROLIN,H.FREITAG,I.C.HEDGE,P.UOTILA,
AND P. G. WILSON. 1993. Chenopodiaceae. In K. Kubitzki [ed.],
Families and genera of vascular plants, 253–281. Springer-Verlag,
Berlin, Germany.
LO
¨HNE, C., AND T. BORSCH. 2005. Phylogenetic utility and molecular
evolution of the petD group II intron in basal angiosperms. Molecular
Biology and Evolution 22: 317–332.
MADDISON, W. P., AND D. R. MADDISON. 2006. Mesquite: a modular system
for evolutionary analysis, version 1.12. Website http://
mesquiteproject.org [accessed 7 October 2007].
MALLIGSON, F. 1922. Serodiagnostische Untersuchungen u¨ber die
Verwandtschaften innerhalb des Centrospermen-Astes des Pflanzen-
reiches. Botanische Archiv 1: 2–20.
MANHART, J. R., AND J. H. RETTIG. 1994. Gene sequence data. In H. D.
Behnke and T. J. Mabry [eds.], Caryophyllales: evolution and
systematics, 235–246. Springer Verlag, Berlin, Germany.
MONSON, R. K., AND S. RAWSTHORNE. 2000. CO
2
assimilation in C
3
–C
4
intermediate plants. In R. C. Leegod, T. D. Sharkey, and S. C. von
Caemmerer [eds.], Photosynthesis: physiology and metabolism, 533–
550. Kluwer, Dordrecht, Netherlands.
MONSON, R. K., J. A. TEERI,M.S.B.KU,J.GUREVITCH,L.J.METS,AND S.
DUDLEY. 1988. Carbon-isotope discrimination by leaves of Flaveria
species exhibiting different amounts of C
3
- and C
4
-cycle co-function.
Planta 174: 145–151.
MUHAIDAT, R., R. F. SAGE,AND N. G. DENGLER. 2007. Diversity of Kranz
anatomy and biochemistry in C
4
eudicots. American Journal of
Botany. 94: 362–381.
MU
¨LLER, K. 2004. PRAP—computation of Bremer support for large data
sets. Molecular Phylogenetics and Evolution 31: 780–782.
MU
¨LLER, K., AND T. BORSCH. 2005a. Phylogenetics of Amaranthaceae
2002 AMERICAN JOURNAL OF BOTANY [Vol. 94
based on matK/trnK sequence data—evidence from parsimony,
likelihood, and Bayesian analyses. Annals of the Missouri Botanical
Garden 92: 66–102.
MU
¨LLER, K., AND T. BORSCH. 2005b. Multiple origins of a unique pollen
feature: stellate pore ornamentation in Amaranthaceae. Grana 44:
266–281.
NIXON, K. C. 1999. The parsimony ratchet, a new method for rapid
parsimony analysis. Cladistics 15: 407–414.
PALMER, J. 1998. A taxonomic revision of Gomphrena (Amaranthaceae) in
Australia. Australian Systematic Botany 11: 73–161.
PEDERSEN, T. M. 1990. Studies in South American Amaranthaceae III
(including one amphi-Atlantic species). Bulletin du Museum national
d’Histoire naturelle, Paris, 4e se´ries 12: 69–97.
PEDERSEN, T. M. 1999. Amaranthaceae. In F. O. Zuloaga and O. Morrone
[eds.], Cata´logo de las plantas vasculares de la Republica Argentina,
16–41. Missouri Botanical Garden Press, St. Louis, Missouri, USA.
PEDERSEN, T. M. 2000. Studies in South American Amaranthaceae. V.
Bonplandia (Corrientes) 10: 83–112.
POSADA, D., AND K. A. CRANDALL. 1998. Modeltest: testing the model of
DNA substitution. Bioinformatics 14: 817–818.
PRATT, D. B. 2003. Phylogeny and morphological evolution of the
Chenopodiaceae-Amaranthaceae alliance. Ph.D dissertation, Iowa
State University, Ames, Iowa, USA.
PYANKOV, V. I., E. G. ARTYUSHEVA,G.E.EDWARDS,C.C.BLACK JR., AND P.
S. SOLTIS. 2001a. Phylogenetic analysis of tribe Salsoleae (Chenopo-
diaceae) based on ribosomal ITS sequences: implications for the
evolution of photosynthesis types. American Journal of Botany 88:
1189–1198.
PYANKOV, V. I., E. V. VOZNESENSKAYA,A.V.KONDRATSCHUK,AND C. C.
BLACK. 1997. A comparative anatomical and biochemical analysis in
Salsola (Chenopodiaceae) species with and without a Kranz type leaf
anatomy: a possible reversion of C
4
to C
3
photosynthesis. American
Journal of Botany 84: 597–606.
PYANKOV, V., H. ZIEGLER,A.KUZ’MIN,AND G. EDWARDS. 2001b. Origin and
evolution of C
4
photosynthesis in the tribe Salsoleae (Chenopodia-
ceae) based on anatomical and biochemical types in leaves and
cotyledons. Plant Systematics and Evolution 230: 43–74.
RAGHAVENDRA, A. S., AND V. S. R. DAS. 1978. Occurrence of C
4
photosynthesis—supplementary list of C
4
plants reported during late
1974–mid 1977. Photosynthetica 12: 200–208.
RAJENDRUDU, G., AND V. S. R. DAS. 1990. C
3
-like carbon isotope
discrimination in C
3
–C
4
intermediate Alternanthera and Parthenium
species. Current Science 59: 377–379.
RAJENDRUDU, G., J. S. R. PRASAD,AND V. S. R. DAS. 1986. C
3
–C
4
Intermediate species in Alternanthera (Amaranthaceae)—leaf anato-
my, CO
2
compensation point, net CO
2
exchange and activities of
photosynthetic enzymes. Plant Physiology 80: 409–414.
ROBERTSON, K. R., AND S. E. CLEMANTS. 2003. Amaranthaceae. In Flora of
North America Editorial Committee [eds.], Flora of North America,
vol. 4, Magnoliophyta: Caryophilidae, part I, 405–456. Oxford
University Press, New York, New York, USA.
RONQUIST, F., AND J. P. HUELSENBECK. 2003. MrBayes 3: Bayesian
phylogenetic inference under mixed models. Bioinformatics 19:
1572–1574.
RUTHSATZ, B., AND U. HOFMANN. 1984. Die verbreitung von C
4
-Pflanzen in
den semiariden Anden NW-Aregentiniens mit einem beitrag zur
blattenanatomie ausgewahlter Beispiele. Phytocoenologia 12: 219–
249.
SAGE, R. F. 2004. The evolution of C
4
photosynthesis. New Phytologist
161: 341–370.
SAGE, R. F., M. R. LI,AND R. K. MONSON. 1999. The taxonomic
distribution of C
4
photosynthesis. In R. F. Sage and R. K. Monson
[eds.], C4 plant biology, 551–584. Academic Press, San Diego,
California, USA.
SA
´NCHEZ-DEL PINO, I., AND H. FLORES OLVERA. 2006. Phylogeny of
Tidestromia (Amaranthaceae, Gomphrenoideae) based on morphol-
ogy. Systematic Botany 31: 689–701.
SCHINZ, H. 1893. Amaranthaceae. In A. Engler and K. Prantl [eds.], Die
natu¨ rlichen Pflanzenfamilien, vol. 1a, 91–118. Engelmann, Leipzig.
SCHU
¨TZE, P., H. FREITAG,AND K. WEISING. 2003. An integrated molecular
and morphological study of the subfamily Suaedoideae Ulbr.
(Chenopodiacea). Plant Systematics and Evolution 239: 257–286.
SWOFFORD, D. L. 2002. PAUP*: phylogenetic analysis using parsimony
(*and other methods), version 4.0b10. Sinauer, Sunderland, Massa-
chusetts, USA.
THIV, M., M. THULIN,N.KILIAN,AND H. P. LINDER. 2006. Eritreo-Arabian
affinities of the Socotran flora as revealed from the molecular
phylogeny of Aerva (Amaranthaceae). Systematic Botany 31: 560–570.
TOWNSEND, C. C. 1985. Amaranthaceae. In R. M. Polhill [ed.], Flora of
tropical East Africa, vol. 7, 1–136. A. A. Balkema, Rotterdam,
Netherlands.
TOWNSEND, C. C. 1993. Amaranthaceae. In K. Kubitzki [ed.], The families
and genera of vascular plants, vol. II, Flowering plants: dicotyledons,
70–91. Springer-Verlag, Berlin, Germany.
ULBRICH, E. 1934. Chenopodiaceae. In A. Engler and K. Prantl [eds.], Die
natu¨ rlichen flanzenfamilien, vol. 16c, 379–584. Engelmann, Leipzig,
Germany.
WANG, R. Z. 2003. C
4
plants in the vegetation of Tibet, China: their natural
occurrence and altitude distribution pattern. Photosynthetica: 41: 21–
26.
WELSH, S. L., C. W. CROMPTON,AND S. E. CLEMANTS. 2003. Chenopodia-
ceae. In Flora of North America Editorial Committee [eds.], Flora of
North America, vol. 4, part 1, 258–404. Oxford University Press,
New York, New York, USA.
ZIEGLER, H., K. H. BATANOUNY,N.SANKHLA,O.P.VYAS,AND W. STICHLER.
1981. The photosynthetic pathway types of some desert plants from
India, Saudi Arabia, Egypt, and Iraq. Oecologia 48: 93–99.
December 2007] SAGEETAL.—C
4
PHOTOSYNTHESIS IN THE AMARANTHACEAE 2003
