Nuclear DNA Variation in Diploid and Polyploid Taxa of Larrea (Zygophyllaceae)

Heredity 63 (1989) 321—328 Received 12 April 1989 The Genetical Society of Great Britain

Nuclear DNA variation in diploid and polyploid taxa of Larrea (Zygophyllaceae)

Lidia Poggio,* * Departamentode Ciencias Biológicas, Facultad de Alicia D. Burghardt* and Ciencias Exactas y Naturales, 1428 Buenos Aires, J. H. Hunziker*+ Argentina. tInstitutode Botánica Darwinion. C.C. 22,1642San Isidro, Argentina. A study of nuclear DNA content was made in telophase nuclei (2C) of the root apex of germinating seed in nine populations of the following species and cytotypes of Larrea: L. nitida (2x), L. divaricata (2x), L. cuneifolia (4x) and L. tridentata (2x,4x,6x). There were no significant differences in DNA content per basic monoploid genome among the diploid taxa nor between the latter and the tetraploid, among tetraploids or between tetraploids and the hexaploid. On the other hand, the difference between means was significant when all diploids were compared with the hexaploid cytotype. These results would indicate: (I) Speciation at the diploid level in Larrea has not produced great differences in DNA content per basic genome. This is in contrast with the related genus Bulnesia. (2) In Larrea there is a slight diminution in DNA content per basic genome when there is an increase in ploidy level. (3) Species of Larrea, Bulnesia and Pintoa (Zygophyllaceae) that inhabit the most arid environments are the ones possessing the highest DNA content. (4) This increase is due to an increment in ploidy level in Larrea and an augment of intrachromosomal DNA in Bulnesia and Pintoa.

INTRODUCTION autopolyploidy (Hunziker et aL, 1972, 1978; Yang et a!., 1977). Thegenus Larrea has an interesting disjunct Bennett (1976, 1987) has suggested that inter- amphitropical distribution covering arid and semi- specific variation in DNA content has adaptive arid regions of Argentina, Chile, Bolivia, Peru, significance and is correlated with the environment Mexico and Southwestern United States. It is sub- and the geographical distribution. In the present divided into two taxa: Sect. Larrea, comprising the investigation the 2C DNA content of L. nitida, L. South American diploid (2n =26)multifoliolate diva ricata, L. cuneifolia and the three cytotypes of species L. nitida and L. ameghinoi and Sect. Bi- L. tridentata is studied. An analysis is attempted folium, which includes bifoliolate species such of the variations that might have occurred during as the South American "jarillas" L. divaricata, speciation and the possible correlation with diploid, and L. cuneifolia, tetraploid (2n =52),and adaptation to arid environments. the North American "gobernadora" or "creosote bush" L. tridentata. This last taxon comprises a diploid cytotype (2n =26),occurring in the Chihuahuan desert, a tetraploid (2n =52)in the MATERIALSAND METHODS Sonoran desert, and an hexaploid (2n =78)in the Mohave desert (Hunziker et a!. 1977, 1978). Theseed was usually collected from 25 randomly The three cytotypes have almost exclusively an chosen individuals in each population. Rep- allopatric distribution and there is a correlation resentative herbarium specimens are deposited at between the increase in ploidy level and the incre- the herbaria of the Facultad de Ciencias Exactas ment in aridity, the Mohave desert being the most y Naturales, Universidad de Buenos Aires extreme. Tetraploid and hexaploid populations (BACFC), the Missouri Botanical Garden (MO) would have originated through interracial and the Instituto de Botánica Darwinion (SI). 322 L. POGGIO, A. D. BURGHARDT AND J. K. HUNZIKER

Abbreviations correspond to J. H. Hunziker (JHH) material was then rinsed three times in SO2 water and A. D. Burghardt (ADB). for 10 minutes each rinse, then rinsed again with The origin of the materials is as follows: distilled water (5 to 15 minutes) and, finally, L. nitida. Argentina, Prov. Mendoza, Depto. Las squashed in 45 per cent acetic acid. The coverslip Heras, Between Punta de Vacas and Arroyo El was removed after freezing with CO2 and the Tambillito, 2200 m elevation. JHH 9870. material was dehydrated in absolute alcohol and L. divaricata. Argentina, Prov. RIo Negro, Depto. mounted in euparal. San Antonio, Las Grutas, 20m elevation. ADB The amount of Feulgen staining per nucleus, 31. Argentina, Prov. La Pampa, Depto. Caleu expressed in arbitrary units, was measured at a Caleu, Anzoategui, 100 m elevation. ADB 5. wavelength of 570 mm using the scanning method L. cunefo1ia. Argentina, Prov. La Rioja, Depto. in a Zeiss Cytoscan. The basic genome DNA con- Capital, Near the airport and RIo de la tent expressed in pg was calculated using A/hum Rodadera, 500 m elevation. JHH 9723. cepa as a standard (2C=33.Spg; Bennett and Argentina, Prov. RIo Negro, Depto. San Smith, 1976). The differences in DNA content Antonio, Route 251, 9 Km North of junction of between populations and species were tested routes 3 and 251. ADB 52. through an analysis of variance and comparisons L, tridentata. (2x). U.S.A., Arizona, Mountview., between means using the Games and Howell Ca.25miles SE of Tucson, 3550 ft. Col. T. W. method (Sokal and Rohlf, 1981) or Scheffé's Yang. method (Scheflé, 1959). L. tridentata. (4x). U.S.A. Arizona, Tucson, Belle- Chromosome studies were based on squashes vue, 2490 ft. Col. T. W. Yang. from root tips pretreated with saturated aqueous L. tridentata. (6x). U.S.A., California, San Bernar- solution of paradichlorobenzene and fixed in a 3 dino Co., Victorville, Vacant lot, 200-300 m from ethyl alcohol: 1 acetic acid solution. After macer- hotel. JHH 10025. U.S.A., Nevada, Nye Co., ation in SN HCI at 20°C; 2 per cent acetic route 95, between Tonopah and Scotty's junc- haematoxylin was used for staining. tion, 52 miles South of Tonopah. JHH 10026. DNA content was measured in telophase nuclei (2C) at the root apex of germinating seed. For RESULTS germination, seeds were placed in petri dishes on wet filter paper. Roots of 05-1 cm length were Chromosomenumbers were determined for all fixed in 3 absolute ethanol: 1 acetic acid during strains. L. divaricata and L. cunefolia were found 1-4 days. A/hum cepa roots were used as a stan- to be diploid and tetraploid as expected and this dard. After fixation, the roots were rinsed 30 confirmed several counts from different regions of minutes in distilled water. Hydrolysis was carried their range made previously on these species (Hun- out with 5 N HCI at 20°C (Deitch et a!., 1968; Fox, ziker et a!., 1978). The chromosome numbers 1969). Different times of hydrolysis were investi- obtained in L. tridentata agree with the innumer- gated and the optimum period determined was 50 able cytological determinations made by Yang minutes. After hydrolysis, the material was rinsed (1967, 1968, 1970) and confirms the diploid, tetra- three times with distilled water for 10 or 15 minutes. ploid and hexaploid nature of the Chihuahuan, Staining was carried out with Feulgen stain at Sonoran and Mohave cytotypes, respectively pH 22 for 2 hours (Teoh and Rees, 1976). The (Yang et a!., 1977).

Table I NuclearDNA content in different populations of Larrea divaricata (2x), L. cuneifolia (4x) and L. tridentata (6x) DNA Content (2C) No. of Species Population (AU.) nuclei G & H(1) L.divaricata ADHn03I 1416±0067 24 2.946NSa (2x) ADB 05 1728±0082 17 L. cuneifolia ADBn052 2217±0115 14 1171 NSh (4x) JHH n°9723 25230227 16 L. tridentata ,IHHn°10025 3042±0120 34 1832 NS5 (6c) JHHn°10026 3'541±0'244 12 Level of significance: a 0Ola; = (1)Games and Howell method, program MCHETV, Sokal and Rohlf (1981). NUCLEAR DNA VARIATION IN LARREA 323

Table 2Nuclear DNA content in populations of species of Larrea DNA content DNA perbasic (2C) genome

Species or Ploidy No. of SE. cytotype level 2n nuclei (A.U.) pg. AU. pg. L. nitida 2x 26 60 1557±0090 1625 0779 0813 L. divaricata 2x 26 41 1545±0056 1614 0773 0807 L. tridentata 2x 26 38 1439± 0086 1502 0719 0751 L. tridentata 4x 52 60 2436±0072 2543 0609 0636 L. cuneifolia 4x 52 30 2380±0 136 2485 0595 0621 L. tridentata 6x 78 46 3173±0113 3313 0529 0552

DNA contents in arbitrary units and in nificantly with ploidy level (y=0.64+043x; F= picograms are summarized in tables I and 2. Figure 9862; P< 10—10) but the estimated regression line I shows the frequency histograms of the species has a gentler slope than a calculated hypothetical and cytotypes indicating the number of nuclei line, which assumes that when the number of studied and their DNA content in arbitrary units. genomes increases, DNA is added as an exact For each population 2 or 3 replicates were made multiple of the DNA content per basic genome which presented nonsignificant differences. (fig. 2(A)). This difference is correlated with the In L. divaricata (diploid), L. cuneifolia (tetra- obtention of a negative regression when DNA con- ploid) and L. tridentata (hexaploid) two popula- tent per basic genome is plotted against ploidy tions from different localities were studied (table level (y=081-0047x; F=9776; P<10'; fig. 1. In the three species, using an approximate test 2(B)). of equality of means (Games and Howell method, In fig. 3 the 2C DNA content is plotted against Sokal and Rohlf, 1981) the differences between ploidy level for the species of Larrea studied here averages of different populations of each species and for the related genera Bulnesia (Poggio and were non significant. As a consequence data were Hunziker, 1986) and Pintoa (Poggio and Naranjo, pooled for each species in table 2. in press). An analysis of variance for the DNA content per basic genome of the species listed in table 2 indicates that there are significant differences DISCUSSION among the taxa (F=9776, P<107). Com- parisons made through Scheffé's method indicated Withinthe genus Larrea the multifoliolate group that there are no significant differences in DNA (Sect. Larrea) is the most primitive and comprises content per basic genome either among diploids the two diploid species L. ameghinoi and L. nitida, or between the tetraploids L. cunefo1ia and L. having 3-16 leaflets. The supposedly more tridentata. There are no significant differences advanced group Sect. Bifolium consists of bifolio- either when the following pairwise comparisons late (L. divaricata (2n =26),L. cuneifolia (2n 52) are made: (1) L. tridentata (diploid) with L. triden- and L. tridentata (2n =26,52, 78)) which have tata (tetraploid) and (2) L. tridentata (tetraploid) their leaves reduced to only one pair of leaflets. with L. tridentata (hexaploid). On the other hand, The bifoliolate species seem to represent a reduc- the difference between means was significant when tional trend in response to aridity (Hunziker et a!., each diploid was compared separately with the 1977). hexaploid cytotype of L. tridentata. These results The DNA 2C contents of L. nitida and L. would indicate that there is a slight diminution in divaricata (table 2), both diploid, but one from DNA content per basic genome when there is an the primitive section and the other from the more increase in ploidy level. advanced, respectively, do not differ significantly. For evaluating the variation in DNA content, L. divaricata has a much wider distribution (Hun.- 2C and per basic genome, in relation to ploidy ziker et a!., 1977, figs. 2.3 and 2.1) and has a wider level, a regression variance analysis was made. In range of habitats. L. nitida is restricted to the both cases linear regression was highly significant coolest and more humid habitats of the whole and the results that were obtained are represented range occupied by L. divaricata. DNA content in in fig. 2. The 2C DNA content increases sig- Patagonian L. ameghinoi was not studied because 324 L. POGGIO, A. D. BURGHARDT AND J. H. HUNZIKER

I I I I I + L.-tdCv1,tata 2) 10- Th + 20 L. d-ivwzcicata

I—

20 L. vUdct

10 C-,

z I I 0 20 + L.wioZiji z 10

I- + L.dc.nta-tct(4x) 20

I L

20. + L.Ldcvt-ta.tct(6x)

: I 0.2 0.8 1 . 2.0Li1}2.6 3.2 3.8 5.0 DNA CONTENT (A.U.)

Figure 1 Frequency histograms: 2C DNA content in arbitrary Units (AU.) plotted against number of nuclei.

of lack of viable seed but it is likely that its content and physiological variations producing adaptation does not differ greatly from that found in L. nitida to aridity and other rigorous conditions (low tem- since their natural hybrid is completely fertile, peratures, wind, etc) in the diploid species of shows normal meiosis with good pairing, without Larrea are apparently not correlated with great noticeable heteromorphic bivalents and both taxa variation of their DNA content. We may conclude could be considered partially sympatric semi- that, speciation at the diploid level in Larrea has species (Hunziker et a!., 1978). The morphological not produced great differences in DNA content NUCLEAR DNA VARIATION IN LARREA 325

4 // z0 7 I/ 3 / U z / I- y=O806—0047x z / F=97760** 0 P<1o / y=O642+O429x zI- z F=98622 ** 0 1— P< 1O-° z rlc-) 0 / U z 2x 4x 6x 0 2x 4x 6x A PLOIDYLEVEL B PLOIDYLEVEL

Figure 2 A =Regression2C total DNA Content on ploidy level in cytotypes of Larrea. Solid line =observed,broken =expected regression. B DNA Content per basic genome plotted against ploidy level (regression line). per basic genome. The opposite has occurred in latter species is an allopolyploid, one of its parental the closely related genus Bulnesia (Poggio and diploids being L. divaricata and the other a now Hunziker, 1986; Poggio et aL,1986). extinct and unknown diploid (Hunziker et at., On the other hand, when the DNA contents 1978). per basic genome of tetraploid L. cunefo1ia and The similarity of DNA content in these auto tetraploid and hexaploid cytotypes of L. tridentata and allotetraploid taxa would indicate either that are analyzed a slight tendency towards diminution DNA per genome has decreased in like manner at is observed when there is an increase in polyploidy the tetraploid level or that some of the ancestral (table 2, fig. 2(B)). diploids had lower DNA content than present day Even when a stronger compaction of DNA in diploids. polyploid nuclei produces an underestimate of the Two alternative explanations have been offered measurements (Verma and Rees, 1974; Kenton, to explain why polyploid taxa may have less DNA. 1984b) it has been observed in several cases that per basic genome than the diploids. According to polyploids have smaller chromosomes and lower one, the polyploids could have been originated by DNA content than the expected (Darlington, 1965; diploids, perhaps extinct at present, that possessed Grant, 1976, 1987; Bennett, 1987; MartInez and lower DNA content. Ginzo, 1985; Poggio and Hunziker, 1986; Poggio These diploids, having the smallest chromo- and Naranjo, in press). In Larrea due to the very somes, would have been preadapted to give rise small size of its chromosomes it is not known to polyploids (Darlington, 1956). whether the diminution affects all chromosomes Another explanation is that saltatory changes or only part of the complement. DNA content per occurred in the genome (DNA reduction) and this basic genome in the diploid taxa of Larrea (0.75- could have happened at the diploid or polyploid O81 pg, see table 2) is slightly higher than the level (Grant, 1987). content of the species that Palacios and Hunziker The fact that interracial autotetraploids of L. (1984) considered as primitive in the closely related trident ata have less DNA content per basic genome genus Bulnesia (B. schickendantzii and B. foliosa, than the related diploids (L. tridentata, L. O58 pg and O5O pg, respectively; Poggio and Hun- divaricata) suggests that the elimination of DNA ziker, 1986). In Bulnesia octoploid B. bonariensis has occurred in the polyploids and not in their has the smallest DNA content per basic genome ancestral diploids. It could be postulated that at and the smallest chromosomes (Poggio and Hun- polyploid level, the DNA elimination leads to a ziker, 1986; Poggio et aL, 1986). more adequate balance between total DNA con- The DNA content of autotetraploid L. triden- tent and a certain cellular parameter. Moreover, tata does not differ from that of L. cunefo1ia. The at the polyploid level, with genetic material dupli- 326 L. POGGIO, A. D. BURGHARDT AND J. H. HUNZIKER

oP.chi.

5.0 -

°B.i'et.

3.5 *L.tri. CD

3.0 Lii oB.ch.. I- Q L) *L.tri. 2.5 L.cun. B.hon. 0

2.0

*L.nit. *L.dtv. 1.5 *L.tri. L.tri.) 0B.8ch. (*Ln) (*L.tri.) .0. °B.fol. °B arb. oB. sar. 0 0.5-

I I I 2x 6x 8x

PLOIDY LEVEL

Figure3 2C total DNA content plotted against ploidy level in cytotypes and species of Bulnesia(fromPoggio and Hunziker, 1986),Pinloa(from Poggio and Naranjo, in press) and Larrea. Between parenthesis the DNA content per diploid genome. Abbreviationsare as follows: Bulnesia B. arh. B.arborea; B. bon. =B.honariensis; B. chi B.chilensis; B. fol. =B.foliosa; B. ret. =B.retarna; B. sar. =B.sarmientoi; B. sch. =B.schickendantzii. Larrea =L.cun. =L.cuneifolia; L. div. =L.divaricata; L. nit. =L.nitida; L. tn. =L.trjdentata. Pinloa =P.chi.P. chilensis. NUCLEAR DNA VARIATION IN LARREA 327 cated, the partial elimination of DNA material is BENNETT. M. D. AND SMITH, J.B.1976.Nuclear DNA amount more easily tolerated. in Angiosperms. Proc. R. Soc. London B, 274, 227-274. DARLINOTON,c.D.1956.Chromosome Botany. G. Allen and In L. tridentata the ploidy level and therefore Unwin Ltd. London. the total DNA content increases towards the west DARLINGTON, C. D. 1965. Cytology. Churchill, Ltd. London. with a concomitant increase in aridity (Hunziker DEITCH, A. D., WAGNER, D. AND RICHARD, R. M. 1968. Condi- et al., 1977, 1978). If total DNA content of tions influencing the intensity of the Feulgen reaction. 3. Zygophyllaceae of the genera Larrea, Bulnesia and Histochem. Cytochem., 16, 371-379. FOX,D. p.1969. Some characteristics in the cold hydrolysis Pintoa is compared (fig. 3) it can be seen than technique for staining plant tissues by the Feulgen reaction. those species that inhabit the most arid environ- J. Histochem. Cvtochem., 17, 266—272. ments (B. retama, B. chilensis, P. chilensis, L. GRANT,W. F.1976. The evolution of karyotype and polyploidy tridentata 6x) are the ones possessing highest DNA in arboreal plants. Taxon, 25, 75-84. GRANT, W. F. 1987. Genome differentiation in higher plants. content. This increase is due to an increment in Chapter I. Urbanska, K. M. (ed.) DIfferentiation Patterns ploidy level in Larrea and an augment of intra- in Higher Plants. Academic Press. London and N. York, chromosomal DNA in Bulnesia and Pintoa (Pog- pp. 9-32. HuNZIKER,J.H.1975.On the geographical origin of Larrea gio and Hunziker, 1986; Poggio and Naranjo, in divaricata (Zygophyllaceae). Ann. Missouri Bot. Gard., 62, press). 497—500. It is possible that changes in DNA content can HUNZIKER, J. H., PALACIOS, R.A., VALES!, A. G. DE AND influence the ecological properties of species. POGGIO, L. 1972. Species disjunctions in Larrea: evidence Increased repetitive DNA may play an important from morphology, cytogenetics, phenolic compounds, and seed albumins. Ann. Missouri Bot. Gard., 59, 224-233. role in gene regulation while gene duplication may HUNZIKER, J. H., PALACIOS, R. A., P00010, L., NARANJO, C. A. increase ecological amplitude because of fixation AND YANG, T. w. 1977. Geographic Distribution, Mor- of different alleles at loci having equal function or phology, Hybridization and Evolution. In Mabry, T. .1., through the acquisition of new genic functions, Hunziker, J. H. and Difeo, 0. R. (eds.) Creosote Bush. (Levin and Funderberg, 1979). It is interesting to Biology and Chemistry of Larrea New world Desert. US/IBP synthesis Series No 6. Hutchinson and Ross, Inc. point out, however, that in both cases (increment Stroudsburg, Penn. pp. 10-47. in ploidy level and intrachromosomal DNA) HUNZ!KER, J. H., PALACIOS, R. A., VALESI, A. 0. DE AND nucleotypic parameters are altered whose P00010, L. 1978. Hybridization in Larrea expression is more a function of total DNA content (Zygophyllaceae): a morphological cytogenetic and than the informational content (Bennett, 1987). chemosystematic study. Bol. Acad. Ciencias Córdoba., 52, 281—3 14. Moreover, if DNA variation is correlated with KENTON, A.1984a.Chromosome evolution in the Gibasis morphological or ecological diversity it would seen linearis group. (Commelinaceae) Ill. Chromosoma (Bert), likely to be adaptive (Kenton, 1984a). The results 90, 303-310. obtained in Bulnesia, Pin toa and Larrea suggest KENTON, A. 1984b. Use of pollen tetrads for routine DNA that there is a correlation between environmental densitometry. Heredity, 53, 667-675. L,EVIN, D. A. AND FUNDERBURG, S.w.1979. Genome size in factors and total DNA content and that the latter Angiosperms: temperate versus tropical species. Amer. may have adaptive significance. Nat. 114, 784—795. MARTINEZ, A. AND GINZO, H. D. 1985. DNA content in Trades- cantia. Can. J. Genet. Cytol., 27, 766—775. Acknowledgements Theauthorswish to express their gratitude PALACIOS, K. A. AND HUNZIKER, J. H. 1984. Revision to Dr. Tien Wei Yang for furnishing seed of the diploid and taxonómica del género Bulnesia. (Zygophyllaceae). Dar- tetraploid cytotypes and to Lic. Beatriz Gonzalez for statistical winiana, 25, 299-320. advice. They also would like to thank the "Comisión Nacional P00010, L. AND HUNZIKER. J. H. 1986. Nuclear DNA content de Energia Atómica" and the "Consejo Nacional de Inves- variation in Bulnesia. J. Hered., 77 43—48. tigaciones Cientificas y Técnicas", both of Argentina for the P00010, L., WULEF, A. F. AND HUNZIKER, .H.1986. Chromo- use of the microdensitometer and for several grants, respec- some size, nuclear volume and DNA content in Bulnesia (Zygophyllaceae). Darwiniana, 27, 25-38. tively. P00Gb, L. AND NARANJO, C. A. Contenido de ADN y evolución en plantas superiores, Acad. Nac. Cienc. Exact., FIs, y Nat. Buenos Aires (In r,ress). SCHF.FFE, H. 1959. The Analysis of Variance. John Wiley, New REFERENCES York. SOKAL. K. R. ANDROHLF,F. J. 1981. The Principles and Practice BENNETT, M. D. 1976.DNA amount, latitude and crop plant of Statistics in Biological Research. Freeman and Co. San distribution. In Jones, K. and Brandham, P. (eds.) Current Francisco. Chromosome Research, Elsevier North-Holland Biomedi- TEOH, S. B. AN!) REES, H. 1976. Nuclear DNA amounts in cal Press, Amsterdam, pp. 151—158. populations of Picea and Pinus species. Heredity, 36,123- BENNETT,M.D. 1987. Variation in genomic form in plants and 137. its ecological implications. New PhytoL 106 (Suppl), 177- VERMA, S. C. AND REES, n. 1974. Nuclear DNA and the evol- 200. ution of allotetraploid Brassicae. Heredity, 33, 61—68. 328 L. POGGO, A. D. BURGHARDT AND J. H. HUNZIKER

YANG, T. w.1967. Chromosome numbers in populations of YANG, T.w. 1970. Major chromosome races of Larrea divaricata creosotebush (Larrea divaricata) in the Chihuahuan and in North America. J. Arizona Acad. Sci, 6, 41-45. Sonoran subdivisions of the North American Desert. J. YANG, T. W.,HUNZIKER.J. H., POGGIO, L. ANDNARANJO, C. Arizona Acad. Sci., 4, 183-184. A.1977. Hybridization between South American "Jarilla" YANG,T.w. 1968. A new chromosome race of Larrea divaricata and North American diploid Creosote bush (Larrea, in Arizona. Western Reserve Acad. Nat. list. Museum. Zygophyllaceae). Plant Syst. EvoL, 126, 33 1-346. Spec. PubI., 2,1-4.