Note to users. If you're seeing this message, it means that your browser cannot find this page's style/presentation instructions -- or possibly that you are using a browser that does not support current Web standards. Find out more about why this message is appearing, and what you can do to make your experience of our site the best it can be.
Functional Divergence of Former Alleles in an Ancient Asexual Invertebrate
Natalia N. Pouchkina-Stantcheva,1*Brian M. McGee,1Chiara Boschetti,1Dimitri Tolleter,2Sohini Chakrabortee,1Antoaneta V. Popova,3Filip Meersman,4David Macherel,2Dirk K. Hincha,3Alan Tunnacliffe1
Theory suggests it should be difficult for asexual organismsto adapt to a changing environment because genetic diversitycan only arise from mutations accumulating within direct antecedentsand not through sexual exchange. In an asexual microinvertebrate,the bdelloid rotifer, we have observed a mechanism by whichsuch organisms could acquire the diversity needed for adaptation.Gene copies most likely representing former alleles have divergedin function so that the proteins they encode play complementaryroles in survival of dry conditions. One protein prevents desiccation-sensitiveenzymes from aggregating during drying, whereas its counterpartdoes not have this activity, but is able to associate with phospholipidbilayers and is potentially involved in maintenance of membraneintegrity. The functional divergence of former alleles observedhere suggests that adoption of asexual reproduction could itselfbe an evolutionary mechanism for the generation of diversity.
1 Institute of Biotechnology, University of Cambridge, Tennis Court Road, Cambridge CB2 1QT, UK. 2 UMR 1191 Physiologie Moléculaire des Semences, Université d'Angers/INH/INRA, 49045 Angers, France. 3 Max-Planck-Institute of Molecular Plant Physiology, Am Mühlenberg 1, D-14476 Potsdam, Germany. 4 Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge CB2 1EW, UK.
* Present address: Department of Biology and Environmental Sciences,University of Sussex, Brighton BN1 9QG, UK.
Present address: Institute of Biophysics, Bulgarian Academyof Sciences, 1113 Sofia, Bulgaria.
Present address: Department of Chemistry, Katholieke UniversiteitLeuven, Celestijnenlaan 200 F, B-3001 Leuven, Belgium.
To whom correspondence should be addressed. E-mail: at10004{at}biotech.cam.ac.uk
Bdelloid rotifers (Rotifera, Bdelloidea) have survived for tensof millions of years without sexual reproduction and meioticrecombination (1–4). Male bdelloid rotifers have neverbeen observed, and the genetic evidence is consistent with fullyasexual reproduction by thelytoky. Long-lasting asexual lineagesare thought to be rare because their apomictic nature does notallow the accumulation of favorable, or the elimination of detrimental,mutations through genetic exchange (5–7). However, oneconsequence of apomixis is that the sequence homogeneity ofgene copies that previously were alleles in sexual ancestorsis no longer maintained by recombination. This allows the formeralleles to accumulate mutations and become divergent—aphenomenon referred to as the Meselson effect (8). Thus, insexually reproducing monogonont rotifers (Rotifera, Monogononta)alleles differ very little from each other at synonymous sites(by up to 2.4% for hsp82), but corresponding gene copies inindividual bdelloid clones can differ by as much as 49% (1).In principle, this effect should allow independent evolutionof former alleles through which they can acquire different functions.
We looked for evidence of functional divergence among formeralleles in a gene set associated with desiccation tolerancein bdelloid rotifers (9, 10). cDNAs representing 100 dehydration-inducedgenes from the bdelloid rotifer Adineta ricciae were identified,one of which encoded a polypeptide related to the group 3 lateembryogenesis abundant (LEA) proteins characterized in plantseeds. LEA proteins are linked with desiccation tolerance inplants, invertebrates, and microorganisms (11). We identifiedtwo similar but distinct sequences and named them Ar-lea-1Aand Ar-lea-1B. Both genes contain nine small introns (Fig. 1A),although there is a major structural difference in exon 2, whichin Ar-lea-1A contains a 132–base pair (bp) segment withno counterpart in Ar-lea-1B. Aligned coding sequences show 13.5%synonymous site divergence (Ks) over the whole gene. This divergenceis much greater than that observed between alleles of sexualanimals, but is within the range of values observed in bdelloidsfor former allele pairs (1, 3, 4, 8).
Fig. 1. Genomic organization of A. ricciae lea genes. (A) Schematic representation of Ar-lea-1A and Ar-lea-1B genes within 5-kb Dra I fragments. Introns are depicted as numbered boxes (red) below lines (scaled dark gray) indicating the sequenced region containing each gene. Genomic organization outside these regions (light gray) is deduced from Southern hybridization analysis. Ar-lea-1B fragments indicated by labeled bars (blue) correspond to probes used in Southern hybridizations. (B) Southern hybridization of A. ricciae genomic DNA with lea gene probes. Each panel contains genomic restriction digests with Dra I, Dra I/Eco RI, and Dra I/Nde I, respectively. Size marker positions are indicated. (C) A. ricciae karyotype and FISH with lea gene probe. (Left) The 12 chromosomes of A. ricciae in a single mitotic nucleus from an embryo stained with 4', 6'-diamidino-2-phenylindole (DAPI). (Right) Interphase nucleus hybridized at high stringency to Ar-lea-1A probe labeled with Alexa 488. Red (superimposed, false color): fluorescent signals; blue: DAPI-labeled DNA. Scale bar: 2 µm.
[View Larger Version of this Image (34K GIF file)]
To confirm the presence of two lea gene copies in the A. ricciaegenome, Southern hybridization experiments were performed withprobes from both the 5' and 3' ends of Ar-lea-1B, which cross-hybridizeto the corresponding regions of Ar-lea-1A (Fig. 1B). Both genesreside on 5.0-kb Dra I genome fragments, but these could bedistinguished by double digestion with either Eco RI or NdeI; a restriction map of each gene was constructed accordingly(Fig. 1, A and B). As further confirmation of lea gene copynumber, fluorescence in situ hybridization (FISH) was carriedout on A. ricciae embryo nuclei. Cytogenetic analysis shows12 chromosomes in this species (Fig. 1C, left), as in the relatedspecies, A. vaga (12). Hybridization with a fluorescent probecorresponding to the whole of Ar-lea-1A produced two signalsin interphase nuclei, consistent with detection of lea geneson two separate chromosomes (Fig. 1C, right). Our cloning andhybridization data show two related, but divergent, lea geneson different chromosomes in A. ricciae, and we interpret theseto be former alleles that have diverged by the Meselson effect.Other interpretations are possible, for example, that the ancestralbdelloid was the result of a hybridization event between specieswith unusually similar lea genes, and that one copy of one leagene from both parents was subsequently lost. However, the simplestinterpretation, consistent with the current understanding ofbdelloid genome structure and evolution (1, 4, 8), is that thetwo lea genes are divergent former alleles. Recent studies suggestthat bdelloid rotifers have four copies of some genes locatedon separate chromosomes, which may indicate that they are ancestrallytetraploids (1, 2), in which the four copies are two gene pairsthat correspond to two pairs of former alleles (3). We havenot obtained evidence to date for a second lea gene pair, althoughwe cannot rule out that they are present in the genome but toodivergent for us to detect with Ar-lea-1A sequences.
Expression of the lea genes was shown by quantitative polymerasechain reaction (PCR) to increase about sevenfold over 24 hoursof drying (fig. S1 and table S1). A similar pattern of expressionof lea genes during dehydration has been observed in other anhydrobioticinvertebrates (11) and is consistent with a role in desiccationtolerance.
The predicted protein sequences of ArLEA1A and ArLEA1B are verysimilar, differing only at 12 amino acid sites of 376 alignedpositions; ArLEA1A is longer by 44 amino acids because of the132-bp indel in exon 2 (Fig. 2A). Both sequences have at leastfour variants of the loosely conserved 11–amino acid motifcharacteristic of group 3 LEA proteins (11, 13) (Fig. 2A andfig. S2A), although positions 4 and 5 are more likely to beapolar. The ArLEA1A and ArLEA1B proteins both have a 19-residuehydrophobic sequence at the N terminus, revealed by a hydropathyplot (14), and a putative variant endoplasmic reticulum (ER)retention signal, ATEL, at the C terminus (Fig. 2A and fig.S2). This suggests that these proteins are localized to or transportedthrough the ER. Most group 3 LEA proteins are highly hydrophilic,with a mean hydropathy (GRAVY) score of –0.97 [SD 0.30;n = 30; dataset of (15)], but both bdelloid proteins score –0.46,similar to moderately hydrophilic proteins, such as bovine serumalbumin (BSA) (GRAVY: –0.43). This reduced hydrophilicityof the bdelloid LEA proteins is unusual and may impact theirstructure.
Fig. 2. Primary and secondary structure of A. ricciae LEA proteins. (A) Alignment of ArLEA1A and ArLEA1B protein sequences showing repeated 11-oligomer motifs. ArLEA1A is 420 residues long with a (predicted) molecular mass of 44.5 kD, whereas ArLEA1B extends for 376 residues with a (predicted) molecular mass of 39.8 kD. Near canonical motifs are orange, green, and yellow; degenerate motifs are gray; highlighted residues differ between the two proteins. The 44-residue indel, shown by dashes, is identical to a more N-terminal sequence whose 11-oligomer motifs are also highlighted orange-yellow-gray-orange. A putative signal peptide is overlined at the N terminus. (B and C) Far-UVCD spectroscopy of ArLEA1A and ArLEA1B in solution and dry state.
[View Larger Version of this Image (42K GIF file)]
Group 3 LEA proteins are largely unstructured in solution, probablybecause their extreme hydrophilicity favors interaction withwater over intrachain binding, but they show increased foldingwhen dried or associated with phospholipid bilayers (11, 16).Secondary structures of recombinant forms of ArLEA1A and ArLEA1B,without putative N-terminal signal peptides, were examined byfar-ultraviolet (far-UV) circular dichroism (CD) spectroscopyin hydrated and dry states. CD spectroscopy of ArLEA1A gavea solution spectrum with a single minimum at 200 nm and lowellipticity at 222 nm, consistent with a disordered structure.However, when dried, its spectrum changed markedly, showingminima near 208 nm and 222 nm, indicative of an -helix (Fig. 2B).In contrast, ArLEA1B has an -helical structure in the hydratedstate, which does not change appreciably on drying (Fig. 2C).Secondary structure content calculated from CD spectra showedthat in ArLEA1A the proportion of -helix increased from 29 to84% on drying, whereas ArLEA1B was 82% -helix in solution, increasingslightly to 87% when dry. Protein denaturation analysis wasperformed by monitoring unfolding at 222 nm on exposure to ureaat a range of concentrations from 0 to 6 M. For a typical globularprotein, unfolding is cooperative and yields a sigmoidal curve.However, structure in ArLEA1B was lost linearly with increasingurea concentration (fig. S2E), which suggests that it existsas a premolten globule without significant tertiary structurein solution (17). The relatively small differences in primarystructure of the bdelloid LEA proteins are therefore responsiblefor markedly different secondary structure.
We tested whether the structural differences between ArLEA1Aand ArLEA1B are reflected in functional divergence. LEA proteinspreserve the activity of desiccation-sensitive enzymes duringdrying (18–20), at least partly through prevention ofaggregation, in what is called molecular shield activity (21).We investigated the ability of both bdelloid LEA proteins tobehave as molecular shields by inhibiting desiccation-inducedaggregation of citrate synthase (CS).
When subjected to drying and rehydration, CS partially denaturesand forms particulate aggregates; however, when dried in thepresence of a group 3 LEA protein, such as AavLEA1 from thenematode Aphelenchus avenae (22), CS aggregation is suppressed(Fig. 3). Other proteins, such as BSA, are not effective. ArLEA1Awas found to reduce CS aggregation as expected, although toa lesser extent than AavLEA1, perhaps because of the lower hydrophilicityof ArLEA1A compared with the nematode protein. However, ArLEA1Bbehaved differently, and drying of CS in its presence resultedin increased aggregation compared with CS dried alone. Indeed,ArLEA1B itself is prone to aggregation (Fig. 3), which ArLEA1Aand AavLEA1 are not, possibly because of its more structurednature. Thus, ArLEA1A shows molecular shield activity in commonwith other group 3 LEA proteins, but ArLEA1B does not and isitself sensitive to desiccation.
Fig. 3. Bdelloid LEA protein antiaggregation assay. Citrate synthase (CS), with or without LEA proteins or BSA, and the latter proteins alone where indicated, were subjected to two cycles of vacuum drying and rehydration. Light scattering by protein particulates was measured by apparent absorption at 340 nm in the spectrophotometer. Error bars show standard deviation (n = 3); ns, not significantly different (P > 0.05); significant values *P < 0.05 or **P < 0.001.
[View Larger Version of this Image (17K GIF file)]
Some LEA proteins have a capacity to associate with and stabilizephospholipid bilayers on dehydration (11, 16, 23). Membraneinteraction was assessed with Fourier transform infrared spectroscopyof liposomes dried in the presence of the bdelloid LEA proteinsor AavLEA1. The gel-to–liquid crystalline phase-transitiontemperature (Tm) of dried palmitoyl oleoyl phosphatidylcholine(POPC) vesicles (59.8° ± 1.2°C) was not affectedby the presence of ArLEA1A (58.2° ± 1.1°C) orAavLEA1 (61.9° ± 5.3°C). However, ArLEA1B significantlydecreased Tm to 51.8° ± 2.9°C, which indicatesthat it interacts with lipids. Further examination of the spectrain the asymmetric phosphate-stretching region revealed a distincteffect of ArLEA1B with a marked shoulder at 1242 cm–1(Fig. 4). The peaks were resolved into two components attributedto P=Oasfree (1262 cm–1) and P=Oas H-bonded (1242 cm–1)(24), similar to the effect of water and sugar (25). The correlationcoefficients for the fitted curves were higher than 0.999. Thesmall bonded P=O population in the absence of protein is becauseof interlipid charge-pair interactions between P=O and cholinegroups, whereas the separation of the two P=O populations isprobably because ArLEA1B was only in contact with the outermonolayer of the liposomes (26). Clearly, a greater proportionof P=O groups are H-bonded in the presence of ArLEA1B comparedwith ArLEA1A (42% as opposed to 30%), whereas AavLEA1 has anintermediate value (36%). These results show that ArLEA1B hasa stronger propensity to interact with dry phospholipid membranesthan ArLEA1A and AavLEA1.
Fig. 4. Bdelloid LEA protein membrane association. Infrared spectra of the asymmetric phosphate stretching region of POPC liposomes dried alone or in the presence of ArLEA1A, ArLEA1B, or AavLEA1. Spectra were recorded at 78°C (liquid-crystalline phase). The solid curve comprises both the measured (dots) and fitted absorbance curves. Normalized peaks were fitted into two bands with maxima at 1262 and 1242 cm–1 corresponding respectively to P=Oasfree (short dashes) and P=Oasbound (long dashes).
[View Larger Version of this Image (16K GIF file)]
In summary, the bdelloid LEA proteins, encoded by gene copiesrepresenting former alleles, have different structures and functions.These functional differences are likely to be adaptive, becauseprevention of protein aggregation and protection of cellularmembranes are essential for survival of desiccation (10, 27).The presence of complementary activities in a single gene pairof a desiccation-tolerant bdelloid rotifer illustrates the potentialfor functional diversity resulting from divergence of formeralleles. The process of abandoning sexual reproduction and meiosis,and the resulting sequence homogenization of homologous chromosomes,is similar to genome duplication, which is a major evolutionaryforce (28, 29) that results in orthologous genes evolving relativelyindependently. Similarly, apomixis could drive evolutionarychange by allowing former alleles to diversify in function andmay partly explain how bdelloid rotifers have, without geneticexchange, diversified into almost 400 taxonomic species (30,31).
18. M. T. Sanchez-Ballesta, M. J. Rodrigo, M. T. Lafuente, A. Granell, L. Zacarias, J. Agric. Food Chem.52, 1950 (2004). [CrossRef] [Web of Science] [Medline]
32. We thank J. Mark Welch for advice on FISH and S. Batey for help with CD spectroscopy. Funded by the Biotechnology and Biological Sciences Research Council (S19912, BB/D001307/1 and 02/A2/P/08059), the Leverhulme Trust (F/09717/B), the Isaac Newton Trust (6.20), and Integrin Advanced Biosystems Ltd. A.V.P. would like to thank the Deutsche Forschungsgemeinschaft for a travel grant. Sequences are deposited into GenBank with accession numbers EF554863 through EF554866.
Received for publication 27 April 2007. Accepted for publication 17 August 2007.
The editors suggest the following Related Resources on Science sites:
In Science Magazine
PERSPECTIVES
Matthew Meselson and David Mark Welch (12 October 2007) Science318 (5848), 202.
[DOI: 10.1126/science.1150197] |Summary »|Full Text »|PDF »
THIS ARTICLE HAS BEEN CITED BY OTHER ARTICLES:
An Unusual Intrinsically Disordered Protein from the Model Legume Lotus japonicus Stabilizes Proteins in Vitro.
S. Haaning, S. Radutoiu, S. V. Hoffmann, J. Dittmer, L. Giehm, D. E. Otzen, and J. Stougaard (2008)
J. Biol. Chem.
283, 31142-31152
|Abstract »|Full Text »|PDF »
From the Cover: Extreme polyploidy in a large bacterium.
J. E. Mendell, K. D. Clements, J. H. Choat, and E. R. Angert (2008)
PNAS
105, 6730-6734
|Abstract »|Full Text »|PDF »
Hydrophilic protein associated with desiccation tolerance exhibits broad protein stabilization function.
S. Chakrabortee, C. Boschetti, L. J. Walton, S. Sarkar, D. C. Rubinsztein, and A. Tunnacliffe (2007)
PNAS
104, 18073-18078
|Abstract »|Full Text »|PDF »