HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Randle, C. P. Articles by Wolfe, A. D. Search for Related Content PubMed Articles by Randle, C. P. Articles by Wolfe, A. D. Agricola Articles by Randle, C. P. Articles by Wolfe, A. D.

The evolution and expression of RBCL in holoparasitic sister-genera Harveya and Hyobanche (Orobanchaceae)¹

²Department of Ecology and Evolutionary Biology and the Natural History Museum, University of Kansas, 1200 Sunnyside Ave., Lawrence, Kansas 66045 USA; ³Department of Evolution, Ecology and Organismal Biology, Ohio State University, 318 W. 12th Ave., Columbus, Ohio 43210 USA

Received for publication December 21, 2004. Accepted for publication June 3, 2005.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX LITERATURE CITED

 
The evolution of holoparasitism decreases the adaptive value of genes maintaining the photosynthetic apparatus. These may become pseudogenes through insertion or deletion events resulting in frameshift mutations, or by the evolution of premature stop codons. The holoparasitic sister genera Harveya and Hyobanche have undergone alternate pathways of evolution and expression at the plastid locus rbcL. An open reading frame in all but a single species of Harveya is maintained by purifying selection and is expressed. However, the function of Rubisco in this putative holoparasite is unknown. Conversely, Hyobanche has undergone rbcL pseudogene formation, and comparison of synonymous and nonsynonymous rates of evolution indicates that selection has not played a role in its evolution. This is complicated by the following findings: multiple pseudogene copies of rbcL exist in tissues of Hyobanche, rbcL transcripts also encode pseudogenes, and the large subunit is present in some tissues of Hyobanche. We hypothesize that the rbcL operon is in a state of degradation as may be expected in a holoparasite and is not endogenously expressed. Rather, the large subunit may be taken up from the host plants, and accumulate in tissues as a result of transpiration.

Key Words: Harveya • Hyobanche • Orobanchaceae • parasitic plant • rbcL • Rubisco • South Africa

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX LITERATURE CITED

 
Parasitic plants may be loosely divided into two categories: hemiparasites, which retain photosynthetic capability, and holoparasites, which derive all photosynthates from plant hosts. Holoparasitic plants do not require endogenous photosynthesis, and therefore selection pressure on photosynthesis-related genes is decreased. The expected outcome of this loss of functional constraint on these genes is an increased rate of evolution resulting in amino acid substitutions for protein-coding genes, pseudogene formation through the evolution of premature stop codons and indels resulting in frameshift, and gene deletion.

Studies of the plastid genomes (plastomes) of holoparasites reveal multiple losses of gene function. The chloroplast genome of the holoparasite Epifagus virginiana (Orobanchaceae), one of the first sequenced, has extremely reduced genome size and content (dePamphilis and Palmer, 1990 ; Morden et al., 1991 ; Wolfe et al., 1992a , b ). Similarly, plastome truncation is evident in other holoparasites of Orobanchaceae, such as Conopholis americana (Wimpee et al., 1991 ; Colwell, 1994 ), Orobanche spp. (Thalouarn et al., 1994 ; Lohan and Wolfe, 1998 ), and Lathraea clandestina (Delavault et al., 1996 ), as well as Cuscuta species (Convolvulaceae; Bömmer et al., 1993 ; Freyer et al., 1995 ) and Cytinus (Cytinaceae; Nickrent et al., 1997a , b ). Although no evidence was found for a plastome in Corynaea (Corynaceae) or Hydnora (Hydnoraceae) (Nickrent et al., 1997b ) in Southern blot studies, 16s rDNA sequences have been obtained (Nickrent et al., 1997a ) from these taxa. While highly divergent, these sequences appear to be functional. On the other hand, 16s sequences from Pilostyles (Apodanthaceae) are probably not functional (Nickrent et al., 1997a ). The loss of cpDNA gene function is also evidenced in more recent lineages. For example, in Boschniakia, Hyobanche, and Orobanche (all genera of Orobanchaceae), rbcL has undergone pseudogene formation (Wolfe and dePamphilis, 1998 ).

Although many holoparasitic plants have reduced plastomes, the plastid has retained function in several genera, primarily in the gene expression apparatus. Pseudogene formation or gene deletion has occurred at all photosynthetic and chemorespiratory loci of the Epifagus virginiana plastome. Of the 42 intact genes in the Epifagus plastome, 38 encode ribosomal proteins, rRNA, and tRNA, while four others have unknown function (Wolfe et al., 1992a , b ). Correctly modified transcripts for a number of these genes have been detected in vivo indicating the retention of plastid function in Epifagus (Ems et al., 1995 ). In green plants, chloroplasts function in a number of nonphotosynthetic, biosynthetic pathways. Therefore it is reasonable to expect the holoparasite plastome to retain minimal function even when photosynthesis is no longer necessary (Bungard, 2004 ).

Studies have also shown the retention and expression of photosynthesis-related genes in holoparasitic plastids. In Cuscuta reflexa, psbA and rbcL are well conserved, despite the deletion of the expression-related genes trnL, rpl2, and rpl23 from the plastome (Bömmer et al., 1993 ). Low-level rbcL expression occurs in the holoparasite Lathraea clandestina (Thalouarn et al., 1989 ; Delavault et al., 1996 ) and in the heterotrophic euglenoid Astasia longa, which also has a significantly reduced plastome (Siemeister and Hächtel, 1990 ). A number of other holoparasites retain a functional copy of rbcL in the plastid. Examination of rates of synonymous and nonsynonymous substitution of rbcL indicates evolution under purifying selection for Harveya purpurea, Striga gesnerioides, Orobanche fasciculata, and O. corymbosa (Wolfe and dePamphilis, 1997 , 1998 ; Leebens-Mack and dePamphilis, 2002 ).

Harveya and Hyobanche are recently derived holoparasitic sister-genera (Wolfe and dePamphilis, 1998 ; Wolfe et al., in press) with alternative pathways of rbcL evolution; Harveya maintains an open reading frame (ORF), while multiple mutations appear to have resulted in pseudogene formation in Hyobanche (Wolfe and dePamphilis, 1998 ; Wolfe and Randle, 2001 ). While it is possible that these mutations are repaired posttranscriptionally (as will be discussed later), we use the term "pseudogene" here in the sense of Wolfe and dePamphilis (1998) , including those genes that can be inferred to be nonfunctional by the presence of indels interrupting the reading frame or by the evolution of premature stop codons in the coding sequence. Both plants are reported to lack chlorophyll or fully developed chloroplasts and are thus presumably incapable of photosynthesis (de la Harpe et al., 1980 ). It is logical to assume that the ancestor of these genera was also a holoparasite. Therefore, it is curious that Harveya should maintain a seemingly functional form of rbcL while Hyobanche does not. In this study, the evolution and expression of rbcL was examined to determine the role of selection in the maintenance of an ORF in Harveya and the loss of function through pseudogene formation in Hyobanche.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX LITERATURE CITED

 
PCR product cloning and DNA sequencing
Tissues were collected in the field and desiccated on silica gel. The text of this manuscript refers to Harveya pulchra, H. leucopharynx, and H. silvatica as species. However, these are subsumed into Harveya huttonii in an upcoming taxonomic revision (C. Randle, unpublished manuscript). DNA was extracted from tissues using a modification of Doyle and Doyle's CTAB protocol (1987) and further purified using the Eluquik system (Schleicher and Schull, Keene, New Hampshire, USA). Amplifications of the entire rbcL operon encompassing the 5'-untranslated region (UTR), rbcL, and the 3'-UTR were carried out using the primers 766+970 and H3UTR (Table 1). Amplification reactions of 50 µL contained 0.5–4.0 µL purified DNA, 0.2 mM dNTPs, 0.15 mM MgCl₂, 0.64 µM each primer, 1 U Taq polymerase (Invitrogen, Carlsbad, California, USA), and 1x Taq polymerase amplification buffer. Reactions were carried out under the following temperature conditions: 2 min at 94°C; 35 cycles of 1 min at 94°C, 1 min at 50°C, and 2 min at 72°C; and a final extension of 10 min at 72°C. Amplified DNA was purified on polyethylene glycol, followed by two ethanol precipitations (85% and 100%, respectively) and resuspended in ddH₂O. Cycle sequencing was carried out using the standard ABI Big Dye 2.0 (Applied Biosystems, Foster City, California, USA) protocol and sequences were obtained using an ABI-3100 automated sequencer (Applied Biosystems).

DNA sequence analysis
Nucleotide sequences of rbcL and inferred amino acid sequences were compared to sequences from Nicotiana tabacum to investigate pseudogene formation, evidenced by the presence of premature stop codons and/or frameshift mutations. Sequences were aligned in Clustal X (Thompson et al., 1997 ) and manually adjusted using the Se-Al data editor (Rambaut, 1996 ). Closely related genera in Orobanchaceae included as outgroups were Alectra sessiliflora, Buchnera floridana, Cycnium racemosum, and Melasma scabrum. The entire region was used to infer the phylogeny of the group using PAUP* version 4.0b10 (Swofford, 2002 ) with parsimony as the optimality criterion. A tree search was carried out with 1000 random sequence additions, five trees held at each step, and tree-bisection-reconnection (TBR) branch swapping. Clade support was estimated with 500 jackknife replicates using TBR branch swapping, 37% deletion, and the "emulate Jac" option. The resulting topology was used in MacClade 3.08a (Maddison and Maddison, 1992 ) to estimate the number of synonymous and nonsynonymous substitutions occurring in rbcL on each branch. To investigate the effects of selection on the evolution of rbcL in Harveya and Hyobanche, the rate of synonymous (d_S) and nonsynonymous substitution (d_N) were calculated using MEGA 2.1 (Kumar et al., 2001 ). Codons for which mutations resulted in premature termination or frameshift mutations (via indels) were excluded from the analysis. Metrics d_S and d_N were calculated in a pairwise manner, using the modified Nei and Gojobori method (Nei and Gojobori, 1986 ; Kumar et al., 2001 ) with Jukes-Cantor correction for the Harveya + Hyobanche lineage and for each genus separately. Values of d_S and d_N were compared in each clade (Harveya + Hyobanche, Harveya, and Hyobanche) using the Wilcoxon signed ranks test in SPSS (SPSS, Inc., 1999 ) to test the null hypothesis of neutral evolution (H_o : d_S = d_N). Three regions of known function in the 5'-UTR were examined for mutations that might preclude expression. The 3'-UTR ends in an inverted repeat (IR), which functions in transcript processing and stability (Schuster and Gruissem, 1991 ; Mayfield et al., 1995 ; Rott et al., 1998 ). Implied RNA sequences were used to infer secondary structure of 3'-IRs on the M-fold server (http://www.bioinfo.rpi.edu/applications/mfold/old/ rna/form1.cgi). The most thermodynamically stable configurations were chosen for comparison, although few differences were evident in less stable configurations. Cloned sequences from Hyobanche were compared to determine if differences existed between them, and parsimony analysis (as before) was used to test the monophyly of sequences from individuals.

Western blot detection of Rubisco
Tissues were collected in the field from nine species of Harveya (H. capensis, H. pulchra, H. purpurea, H. scarlatina, H. silvatica, H. speciosa, H. squamosa, H. stenosiphon, and H. coccinea), four species of Hyobanche (Hy. atropurpurea, Hy. glabrata, Hy. rubra, and Hy. sanguinea) and from five species of closely related hemiparasites (Alectra capensis, Buchnera glabra, Cycnium racemosum, Melasma scabrum, and Pedicularis lanceolata), which served as positive controls. Rubisco may be differentially expressed in plant organs (Mayak et al., 1998 ; McCormac et al., 2001 ) and therefore, scale-leaves, stems, calyces, and corollas were collected from Harveya and Hyobanche populations when feasible (if more than 15 flowering stems were evident in a population). Tissues were flash frozen in liquid nitrogen and transported to the lab on dry ice before storage at –80°C.

Tissues were ground in extraction buffer (100 mM Tris, 0.5% polyethylene glycol, 1 mM EDTA, 100 mM 2-mercaptoethanol; pH 8.0) and centrifuged to remove nonsoluble components. The volume of extraction buffer used was normalized according to dry tissue weight, 1.0 mL extraction buffer per 0.5 g dry tissue. Extracts were combined with loading buffer (100 mM Tris, 4.0% SDS, 2.0% [w/v] bromophenol blue, 20% glycerol, and 200 mM dithiothreotol) and were separated on 10% polyacrylamide gels containing 0.1% SDS. In one lane of each gel, 7.5 µg of Spinacia-purified Rubisco (Sigma, St. Louis, Missouri, USA) was used as a positive control. Full Range Rainbow Molecular Weight Marker (Amersham Pharmacia, Piscataway, New Jersey, USA) was run in one lane on each gel to estimate band size.

Gels were electrophoretically blotted onto Hybond-P membranes (Amersham-Pharmacia). Binding sites on membranes were blocked for 1 h in 5% nonfat dry milk (with 0.1% Tween 20) in phosphate-buffered saline solution (Sambrook et al., 1989 ). Two anti-Rubisco primary antibodies were used to assay for the presence of Rubisco on membranes. Anti-Rubisco serum (Sigma) was raised in rabbits against purified Spinacia Rubisco. A second primary antibody, consisting of the IgY fraction of chicken serum raised against a peptide target conserved in all type I large subunits (LSU), was available commercially (Agrisera, Vännäs, Sweden). Primary antibodies were diluted in 5% nonfat dry milk/phosphate-buffered saline solution, at dilutions of 1 : 1000 for rabbit anti-Rubisco serum and 1 : 5000 for the chicken anti-Rubisco IgY. Membranes were incubated in the diluted primary antibody for 1 h at room temperature, and then probed with a horseradish peroxidase (HRP)-conjugated secondary antibody, either 1 : 5000 goat anti-rabbit IgG-HRP (Amersham Pharmacia) or 1 : 5000 rabbit anti-chicken IgY HRP (Pierce, Rockford, Illinois, USA). Banding patterns were visualized using the ECL Plus Western blotting system (Amersham Pharmacia) and fluorescence images were captured on ECL Hyperfilm (Amersham Pharmacia).

RNA extraction, RT-PCR, and transcript sequencing
RNA was extracted from tissues frozen in liquid nitrogen as described. Corolla tissue was available from two individuals of Hyobanche sanguinea and one of Hy. glabrata. Additionally, RNA was extracted from calyx material from one individual of each of three species of Harveya: H. capensis, H. coccinea and H. stenosiphon for use as positive controls. RNA was also extracted from stems of Hy. glabrata and Hy. sanguinea that had no LSU immunoreactivity on Western blots, to serve as negative controls. Tissues were ground in liquid nitrogen and RNA was purified using the RNeasy for Plant Tissue kit (Qiagen, Valencia, California, USA). DNA was degraded with DNase I (Qiagen) following the "on-column" protocol option. RNAs were separated on 1% agarose to estimate yield and purity.

RNAs were then used to construct cDNA libraries using reverse-transcriptase (RT) PCR. All reagents were purchased from Invitrogen (Carlsbad, California, USA), and final concentrations are given in brackets. RNA extract (5 µL) was added to 1 µL random primers (250 ng/mL), 1 µL 10 mM dNTPs (0.5 mM), and 6 µL ddH₂O. The mixture was incubated at 65°C for 5 min and immediately transferred to ice. After adding 5x first strand buffer (1x), 1 µL 0.1 M DTT (5 mM), 1 µL superscript III reverse-transcriptase (10 U/ µL), and 1 µL RNase-Out recombinant RNase inhibitor (2 U/µL) were added to the mixture, which was incubated at room temperature for 5 min, followed by incubation at 50°C for 45 min. cDNAs were electrophoresed on 1% agarose to visualize RT-PCR products.

Transcripts of rbcL were then amplified from cDNA libraries in 50-µL reactions, using the protocol described. As an additional negative control, for each cDNA reaction, a reaction was carried out using an equivalent amount of purified RNA to insure that amplified fragments represented RNA transcripts rather than DNA contamination of RNA. Sequencing of cDNA amplicons was carried out as indicated before.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX LITERATURE CITED

 
The evolution of the rbcL coding region in Harveya and Hyobanche
The full rbcL operon (the 5'-UTR, rbcL, and the 3'-UTR) ranged in length from 1698 base pairs in Harveya huttonii to 2104 base pairs in Hyobanche rubra (see Appendix for voucher and GenBank accession numbers). All species of Hyobanche examined have premature stop codons and deletions resulting in frameshifts in their rbcL sequences, thus suggesting the evolution of pseudogenes (Fig. 1). All species of Harveya examined maintain intact open reading frames at this locus, except for Harveya huttonii, which bears two autapomorphic deletions, one of four and the other of 341 bases. To insure that these deletions were not the result of PCR error, rbcL was re-amplified and re-sequenced with the same result. Alignment of the combined 5'-UTR, rbcL, and the 3'-UTR sequences for 20 taxa yielded a matrix of 2193 characters, of which 99 were parsimony-informative. A parsimony search of combined sequences resulted in six most parsimonious trees (CI = 0.870; RI = 0.850). Four of the mutations contributing to pseudogene formation are synapomorphies, supporting clades within the Hyobanche lineage, including two deletions uniting Hy. sanguinea, Hy. rubra, and Hy. glabrata, as well as two substitutions resulting in premature stop codons that unite Hy. rubra and Hy. glabrata (Fig. 1).

View larger version (35K): [in this window] [in a new window]   Fig. 1. Phylogram of Harveya and Hyobanche, with outgroups Alectra sessiliflora, Buchnera floridiana, Cycnium racemosum, and Melasma scabrum inferred from the DNA sequence of the rbcL operon composed of the 5'-UTR, rbcL (coding sequence) and the 3'-UTR. This is one of six most parsimonious trees (CI = 0.870; RI = 0.850; 315 steps). Support values from 500 jackknife replicates (10 random sequence additions each) appear above branches. * indicates branches that collapse in strict consensus. () indicates an rbcL pseudogene evidenced by a deletion resulting in frameshifts (tick mark), or the evolution of a premature stop codon (). Deletions resulting in frameshift (base position of the aligned rbcL sequence): (1) 45–386; (2) 1217– 1220; (3) 524; (4) 923–927; (5) 225; (6) 517–523. Substitutions resulting in the evolution of a premature stop codon (base position of the aligned rbcL sequence): (a) 522; (b) 610; (c) 60; (d) 1066

The role of selection in the evolution of rbcL was investigated by comparison of d_S/d_N ratios for each clade. The inclusion of Harveya huttonii (which has a 341 base deletion) requires a significant portion of the matrix to be ignored, and therefore, calculation of d_S and d_N values was performed with and without H. huttonii. In the Harveya lineage d_S/d_N is significantly greater than 1.0 with or without the inclusion of H. huttonii in the analysis (with: d_S/d_N = 11.0, P < 0.001; without: d_S/d_N = 40.50, P < 0.001; Table 2). This indicates that rbcL likely has evolved under purifying selection in the Harveya lineage. In the Hyobanche lineage, rbcL appears to have evolved randomly (without functional constraint) as d_S/d_N is not significantly different from 1.0 (with H. huttonii: d_S/d_N = 0.90, P = 0.686; without: d_S/d_N = 0.86, P = 0.249).

The 5'-UTR contains three sites of known function in rbcL expression: prokaryote-like -35 and -10 transcription promoter sequences, and a ribosome-binding site just upstream of the rbcL start codon (Shinozaki and Sugiura, 1982 ). At two of the three sites of known function of the 5'-UTR, one may infer several changes in Harveya, compared to those of the hemiparasitic taxa and Hyobanche (Fig. 2a). However, the sites at which these changes occur, within the -10 promoter and the ribosome-binding site, are polymorphic in Harveya. The substitution at the third position of the -10 transcription promoter is a synapomorphy uniting the Harveya huttonii + H. coccinea + H. pulchra + H. silvatica clade. The substitution at the fourth position of the ribosome-binding site is autapomorphic for H. stenosiphon. The majority of Harveya species have sequences identical to their photosynthetic relatives at these sites of known function. Therefore, in most species of Harveya, substitutions in the 5'-UTR are not sufficient to preclude rbcL expression. Hyobanche 5'-UTR sequences are identical to those of the hemiparasites at the functional sites.

View larger version (36K): [in this window] [in a new window]   Fig. 2. (A) The 5'-UTR contains three sequences of known function in rbcL expression, the -35 and -10 transcription promoters and the ribosome-binding site, six bases upstream of the rbcL start codon. Harveya is polymorphic at two of these sites, substitutions evident in the -10 transcription promoter and the ribosome-binding site. In both cases, the majority of Harveya species have identical nucleotide sequences as the hemiparasites and Hyobanche. At the fourth position of the -10 transcription promoter, the change from a C to a T is synapomorphic for the clade containing Harveya pulchra, H. huttonii, H. coccinea, and H. silvatica (Figs. 1 and 2). At the second position of the ribosome-binding site, the substitution from G to A is autapomorphic in Harveya stenosiphon. (B) Stem structures of the 3'-UTR-IR (in gray) are nearly identical in all species examined. Sequences here vary mainly by terminal loop length. Harveya sequences fall into three groups varying by loop length: Harveya 1—clade H. huttonii, H. coccinea, H. pulchra, H. silvatica, H. speciosa, and H. scarlatina; Harveya 2—clade H. capensis, H. purpurea, H. bolusii, H. squamosa, and H. stenosiphon; Harveya 3—H. hyobanchoides. The increase in the stem length of Harveya hyobanchoides is probably due to a duplication event of the motif marked in boxes

Western blot detection of Rubisco
Initially, all Western blotting was performed using an antibody raised in rabbits against purified Spinacia Rubisco. Rubisco is a hexadecamer, and subunits dissociate when subject to an electric current, detergents, or freeze–thaw cycles (Ru et al., 2000 ). Therefore, multiple bands were present in some samples, representing different combinations of large and small subunits. Bands representing large and small subunits of Rubisco were present in tissues of both Harveya and Hyobanche, albeit with immunoreactivity 20–100 fold less than in an equivalent quantity of Spinacia leaf tissue. These results were anomalous, given the presence of pseudogenes at the rbcL locus in all specimens of Hyobanche examined. Therefore, a second antibody generated against a peptide fragment conserved in all type I, Rubisco large subunits was used in an attempt to replicate these data. This second assay confirmed previous findings (Fig. 3, Table 3).

View larger version (42K): [in this window] [in a new window]   Fig. 3. Detection of Rubisco by means of western blot. Initially (left), an antibody raised against purified Spinacia Rubisco was used to detect Rubisco in tissues of (A) Harveya capensis leaf extract, (B) Hyobanche glabrata calyx extract, and (C) purified Spinacia Rubisco (7.5 µg total protein). Both the large (LSU) and small subunit (SSU) were present in Harveya and Hyobanche tissues as evidenced by oligomer bands composed of both subunits (160 kDa, 140 kDA, 90 kDa, and 70 kDa). The presence of a 112-kDa band and a 56-kDa band indicate the large subunit dimers and monomers, respectively. A second antibody raised against a peptide conserved in all type I Rubisco confirmed this result (right)

Hibberd et al. (1998) found that Rubisco expression in Cuscuta reflexa was limited to bands of cells adjacent to vascular bundles of the stem. The Western blot analysis presented in this study is too crude to detect Rubisco in very localized tissues. It should be noted therefore that Rubisco may occur in organs in which it was not detected, possibly in very specialized tissues or cells.

In any case, the presence of the LSU in Hyobanche tissues is anomalous given the apparent pseudogene in all four species. Two hypotheses were tested to further investigate this observation. First, it is possible that multiple copies of rbcL exist, at least one of which encodes an open reading frame. This may be the result of haplotype polymorphism, heteroplasmy, or duplication and transfer to another cellular compartment (see Wolfe and Randle, 2004 , for a complete discussion). Alternatively, pseudogene transcripts may be repaired pretranslationally, restoring an open reading frame. Regulatory mechanisms governing mRNA alteration are largely imported from the nucleus, and many nuclear-encoded factors present in the chloroplast are yet to be characterized (Schuster and Gruissem, 1991 ; Herrman et al., 1992 ; Gillham et al., 1994 ; Mayfield et al., 1995 ; Danon, 1997 ). Such modifications include cis- and trans-splicing, intron splicing, and sequence-specific de-amination of cytosine residues resulting in conversion to uracil. To test the multiple-copies hypothesis, rbcL was cloned and sequenced from one individual of each of four species of Hyobanche, and sequences were examined for mutations resulting in pseudogenes. To test weather RNA editing restored the open reading frame, rbcL transcripts were sequenced using RT-PCR.

Multiple copies of rbcL in Hyobanche
In the preliminary experiments, rbcL amplification primers annealing within the 5'- and 3'- untranslated regions were used. For the individual utilized, 20 sequences were obtained. Two haplotypes were recovered that differed by a single base pair. However, when fragments were amplified using primers annealing just within the coding sequence of rbcL, more haplotypes were recovered. Hyobanche glabrata had five unique haplotypes (GenBank accession numbers DQ017801–DQ01785), Hy. sanguinea (DQ017806–DQ017812) had five, Hy. atropurpurea (DQ17813–DQ17818) had six, and Hy. rubra had 11 unique haplotypes (DQ017819–DQ017829). However, each sequence obtained was a pseudogene.

To test the monophyly of haplotypes obtained from individuals, phylogenetic analysis was carried out on the aligned matrix of cloned sequences. Of 1435 characters, 62 were parsimony-informative. One thousand random addition searches resulted in one most parsimonious tree of 112 steps; CI = 0.821, RI = 0.931. Sequences were monophyletic with respect to the individual from which they were obtained with the exception of Hy. atropurpurea sequences (Fig. 4). All but one of the Hy. atropurpurea sequences formed a clade at the base of the tree, which was strongly supported with a jackknife of 99%. The remaining sequence (clone 97.130.1–16) was strongly supported as sister to the Hy. sanguinea clade, with a jackknife value of 91%. To insure that this was not the result of contamination, the transformant colony was isolated from the original plate and re-sequenced, with the same result.

View larger version (26K): [in this window] [in a new window]   Fig. 4. Jackknife consensus of rbcL cloned sequences (all of which were pseudogenes) from four individuals, one of each species of Hyobanche. The aligned matrix consisted of 1435 characters and resulted in one most parsimonious tree (length = 112, CI = 0.821, RI= 0.931). Redundant sequences were excluded from the analysis, and therefore, each terminal represents a distinct haplotype. Sequences from Harveya species were used to infer the root of the tree. Sequences were moderately to strongly supported as monophyletic with regard to the individual from which they were cloned, with the exception of sequences from a single individual of Hyobanche atropurpurea, which were resolved as paraphyletic

The transcript sequence obtained from Hy. glabrata cDNA was identical to one of the copies previously sequenced in Hy. glabrata. Likewise, the sequence obtained from one individual of Hy. sanguinea (r03.03) was nearly identical to another sequence previously obtained from this species, differing by a single autapomorphic base substitution. However, the other sequence obtained from Hy. sanguinea differed strongly from any other haplotype yet sequenced (r03.08). When this sequence was included in phylogenetic analyses of the matrix of all Hyobanche rbcL sequences, it was resolved in a tritomy at the base of the clade containing all other sequences from Hy. glabrata, Hy. sanguinea, Hy. rubra, and Hy. atropurpurea 97.130.1–16 with a jackknife of 99%. This placement did not affect other relationships in the tree. The tree required nine more steps with the inclusion of this cDNA sequence (the length increased from 119 to 128 steps), but the amount of homoplasy did not change drastically (CI = 0.812; RI = 0.947).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX LITERATURE CITED

 
The presence of rbcL pseudogenes in Hyobanche indicates the loss of functional constraint on this locus. This finding was supported by analysis of rates of synonymous and nonsynonymous substitution. Pseudogene formation and nonsynonymous substitution rate increases may have occurred concomitantly with the loss of rbcL function. Conversely, rbcL has evolved under purifying selection in Harveya, as evidenced by open reading frames in all species except H. huttonii, and an overall high ratio of d_S/d_N. Using a novel power analysis of d_S/d_N ratios among H. purpurea, H. capensis, and the closely related hemiparasite Alectra sessiliflora, Leebens-Mack and dePamphilis (2002) demonstrated purifying selection on rbcL in the Harveya lineage, with a low probability of failing to reject a false null hypothesis (H₀ : d_S = d_N).

Changes in the 5'- and 3'-UTRs of Harveya and Hyobanche do not appear sufficient to preclude gene expression. The -35 and -10 promoters and the ribosome binding site of the 5'-UTR are identical in sequence and approximate position in all species of Hyobanche and most species of Harveya. Furthermore, these sites are strongly similar in sequence and position to those in hemiparasitic plants with known photosynthetic capability. The terminal inverted repeat of the 3'-UTR of these holoparasites is similar to that of hemiparasitic relatives and Nicotiana in sequence and inferred structure. This result is similar to that found by Wolfe and dePamphilis (1997) in parasitic plants from the genera Cuscuta and Orobanche, in that the parasites of the same lineage bore either an rbcL open reading frames or a pseudogene. In their study, however, major deletions were detected in parasitic plants upstream of the terminal inverted repeat, a region not investigated here. In the present study, pseudogene formation in rbcL was not accompanied by major changes in the 5'- or 3'- flanking sequences in any instance. However, some aspects of flanking region function may not yet be characterized, and other changes in sequence and structure not examined here may affect rbcL gene expression (Monde et al., 2000 ; Kuroda and Maliga, 2001 ; McCormac et al., 2001 ; Whitney and Andrews, 2001 ).

The most surprising result of this study was the presence of Rubisco in tissues of holoparasites Harveya and Hyobanche. Expression of rbcL in Harveya is a reasonable outcome based on the maintenance of an open reading frame, evidence of purifying selection and apparently functional flanking sequences, but begs the question of rbcL function in a nonphotosynthetic plant. Evidence of Rubisco in Hyobanche is anomalous given the presence of an rbcL pseudogene in more than 30 specimens examined in this study and in others (Wolfe and Randle, 2001 , 2004 ).

Evolution and expression of rbcL in Harveya
Ignoring the conspicuous presence of Rubisco in Hyobanche, the simplest explanation for Rubisco in Harveya is expression of rbcL within the plastid, as indicated by the presence of an rbcL ORF, low rate of evolution at nonsynonymous vs. synonymous sites, and functional RNA transcript sequences. Of course, this raises the question of Rubisco function in a putative holoparasite. Several analogous conditions exist in organisms in which heterotrophy is a derived condition. Astasia longa, a heterotrophic euglenoid, retains an rbcL ORF despite the absence of photosynthetic pigments and an otherwise truncated plastid genome compared to that of photosynthetic euglenoids. Furthermore, rbcL transcripts and the LSU have been detected in Astasia longa (Siemeister and Hächtel, 1990 ). Cuscuta reflexa is a stem holoparasite that exhibits low levels of light-induced ¹⁴CO₂ assimilation, presumably by the action of Rubisco (Machado and Zetsche, 1990 ). Transcripts of rbcL were detected at low concentrations despite the loss of the 3'-UTR-IR and point mutations in the promoters (Haberhausen et al., 1992 ). Lathraea clandestina, another holoparasite, also has an rbcL ORF (Delavault et al., 1995 ) accompanied by the presence and activity of Rubisco in leaf tissue (Bricaud et al., 1986 ; Thalouarn et al., 1989 ) despite significant divergence and reduction of the plastid genome and a lack of photosynthetically competent chloroplasts (Delavault et al., 1996 ). The evolution of rbcL in Lathraea may differ from that in Harveya, as retention of an rbcL reading frame in Lathraea was shown to be stochastic, rather than functionally constrained (Leebens-Mack and dePamphilis, 2002 ). However, one might expect negative selection to act quickly in terminating the expression of genes that do not contribute to fitness, as production of Rubisco may be costly to Lathraea, even at levels lower than photosynthetic plants. A recent loss of photosynthetic ability in Lathraea clandestina may explain this finding, or alternatively, the recent loss of Rubisco function following an older transition to holoparasitism. The same analysis inferred no loss in functional constraint on rbcL in holoparasitic lineages Striga gesneroides and Orobanche fasciculata/O. corymbosa and had sufficient statistical power to reject a false null hypothesis (Leebens-Mack and dePamphilis, 2002 ).

The maintenance of rbcL function in parasitic plants is puzzling. Harveya appears in most ways to be a holoparasite. All species bear strongly reduced leaves. Tissues of H. huttonii and H. squamosa were shown to contain few if any plastids (de la Harpe et al., 1980 ) and, furthermore, these two species have been reported to lack chlorophyll entirely (de la Harpe et al., 1981 ). However, plastids almost certainly exist in cells of Harveya species as they do in more ancient holoparasitic lineages, namely Conopholis, Epifagus, and Orobanche, and as evidenced in this study by the presence of rbcL in Harveya. One may also question the reported lack of chlorophyll, based on the presence of green pigmentation in the leaves and calyces of many live specimens. In this study, Rubisco immunoreactivity was greatest in these tissues of Harveya. The swollen stigma of H. capensis is also conspicuously green against the white background of the corolla tube, indicating that chlorophyll perhaps serves as a pollinator attractant. Further, the corolla limbs of H. hyobanchoides are also of a deep, chlorophyllous green color.

Therefore, one possibility is that Harveya is a cryptic hemiparasite and may carry out photosynthesis at low levels. Conversely, Rubisco may play a nonphotosynthetic role in Harveya.

The role of Rubisco in nonphotosynthetic plants
The expression of Rubisco in nonphotosynthetic plants, indicates the importance of this enzyme in a nonphotosynthetic role. Rubisco may have a similar function in photosynthetic plants that has been overlooked due to the overwhelming importance of fixing CO₂ in photosynthesis. Several suggested functions of Rubisco other than photosynthetic carbon assimilation include the synthesis of serine and glycine through oxygenation in the glycolate pathway (Siemeister and Hächtel, 1990 ; Wolfe and dePamphilis, 1997 ), and the recycling of internal carbon (Thalouarn et al., 1989 ). In nonphotosynthetic plants, Cuscuta europaea was shown to lack Rubisco activity while maintaining higher than average PEP-carboxylase activity, higher even than its photosynthetic relative Ipomoea, while C. reflexa maintained Rubisco activity but not PEP-carboxylase activity (Machado and Zetsche, 1990 ), suggesting perhaps an overlap in function of these enzymes.

PEP-carboxylase fixes CO₂ nonphotosynthetically and may function in tricarboxylic acid intermediate synthesis needed for the synthesis of amino acids and chlorophyll, generation of NADPH, recapture and recycling of CO₂, carbon metabolism in aquatic plants, malate fermentation, cyanide-resistant respiration that prevents the buildup of a large number of sugars, nitrogen assimilation and amino acid synthesis, maintenance of cytoplasmic pH, maintenance of electroneutrality, wavelength-mediated light response, and the amelioration of low temperature sensitivity (Latzko and Kelly, 1983 ). PEP carboxylase activity in hemiparasites Viscum album (Viscaceae), Thesium humifusum, and Osyris alba (Santalaceae) and in holoparasites Orobanche hederae and Lathraea clandestina was shown to be higher than or comparable to the C4 plant Sorghum bicolor (Renaudin et al., 1982 ). This is significant in that C4 plants in general display higher PEP-C activity than C3 plants from which these parasites were presumably derived (Latzko and Kelly, 1983 ). In Platycerium coronarium (staghorn fern), PEP-C and Rubisco activity are co-regulated by concentrations of atmospheric CO₂. Calluses cultured in high CO₂ environments had an increase in PEP-C activity and a decrease in Rubisco activity (Kwa et al., 1997 ). In C3 plants, PEP-C is active in both the cytosol and the chloroplast, but in the holoparasite Lathraea clandestina, PEP-C expression is limited to the cytosol (Renaudin et al., 1984 ). This suggests that Rubisco (which is minimally expressed in Lathraea clandestina) may have taken over chloroplast PEP-C function. Perhaps this change in the expression pathway indicates a change in function. In holoparasites, some of these functions may be particularly important: (1) in synthesis of necessary organic compounds not obtainable from host plants, (2) in the recycling of carbon that may accumulate detrimentally in plants that respire but do not photosynthesize, and (3) maintenance of pH and electroneutrality through oxidation or carboxylation. Alternatively, the products of Rubisco function may be important in maintaining an appropriate osmotic balance between parasite and host (Stewart and Press, 1990 ). Recycling excess CO₂ may be especially important in parasites that spend most of the life cycle in high CO₂ environments, underground as tubercles on host roots.

Evolution and expression of rbcL in Hyobanche
The discovery of Rubisco in tissues of Hyobanche was unexpected in light of the presence of an rbcL pseudogene as documented here and in previous studies (Wolfe and dePamphilis, 1998 ; Wolfe and Randle, 2001 ). While other parasitic plants have been shown to bear an rbcL pseudogene or deletion, few studies have gone so far as to assay for the presence of rbcL transcripts or the large subunit in these plants. Of these, an absence of expression has been reported only in Epifagus virginiana, which bears an rbcL pseudogene resulting from a deletion of a large portion of the gene (dePamphilis and Palmer, 1990 ). Alternatively, rbcL expression was not detected in Orobanche hederae and O. minor, holoparasites bearing an rbcL ORF (Thalouarn et al., 1994 ).

While it remains possible that an intact rbcL open reading frame exists in species of Hyobanche, the present study failed to corroborate this hypothesis. First, no such ORF was discovered in cloning experiments, and second, rbcL transcripts encoded pseudogenes. That multiple copies of rbcL exist in Hyobanche does nothing to clarify the presence of the LSU in tissues of that plant. In a recent study, the holoparasitic species Cynomorium coccineum (Cynomoriaceae) had multiple divergent copies of 23s rDNA (Garcia et al., 2004 ). Such findings indicate that the inference of species trees from gene trees may be invalid when cpDNA are used due to paralogy (see further discussion in Wolfe and Randle, 2004 ).

This second point, that RNA transcripts encoded pseudogenes is more decisive, in that if an ORF were maintained by purifying selection, one would expect the open reading frame to be represented in the pool of RNA transcripts. Furthermore, these results do not support the second hypothesis, that rbcL pseudogenes are repaired by means of posttranscriptional modification.

The case for endogenous expression of rbcL in Harveya seems clear cut. The locus appears to be evolving under selective pressure, is transcribed, and the large subunit of Rubisco is present in at least some tissues. Only the function of Rubisco in Harveya remains in question. On the contrary, nearly every aspect of the evolution and expression of rbcL in Hyobanche indicates a lineage in which holoparasitism is plesiomorphic; gene systems that support photosynthesis are in an early stage of degradation. This is contradicted only by the presence of the end product of expression, the large subunit of Rubisco. The presence of multiple copies of the gene and pseudogene transcripts, while bizarre and interesting, does not allow any insight into how the large subunit of Rubisco might be expressed. Therefore, we present an alternate hypothesis, which will not be tested in the current study; namely, that rbcL is not expressed endogenously, but obtained from host plants.

Exogenous acquisition of Rubisco from hosts of Hyobanche
The evolution of haustoria in Orobanchaceae is complex and may include convergence and reversal according to most recent phylogenetic hypotheses (Young et al., 1999 ). To further complicate matters, Hyobanche displays a unique type of leaf haustorium not found in any other species of parasitic plant, with the possible exception of Orobanche teucrii (Weber, 1980 ), certainly an example of parallel evolution. Initially, Hyobanche seedlings form a primary haustorium at the base of the emerging radical from which rhizomes grow. In later stages, the primary haustorium disappears altogether, and all nutrients are obtained from the host by means of secondary haustoria, which arise from the scale leaves of rhizomes (Kuijt et al., 1978 ; Weber and Visser, 1980 ). Conversely, Harveya species have been reported as having primary haustoria. If secondary haustoria are present, which has not been demonstrated thoroughly, they arise from lateral roots.

Several studies have shown the activity of cell-wall- and membrane-degrading enzymes (cellulases, proteases, pectin methylesterases, and polygalacturonases) at the interface between metabolically active haustorial parenchyma and cells of the host root in Orobanchaceae (Singh and Singh, 1993 ; Losner-Goshen et al., 1998 ). Most of the transport of carbon from the host appears to be apoplastic, but Striga hermonthica may have xylem carbon concentrations five times greater than its host, Sorghum bicolor, although amino acid content in Striga leaf tissue mirrors that of its host (Stewart and Press, 1990 ). Light microscopy studies of cellular markers in primary haustoria of O. crenata and host Vicia narbonensis show phloem– phloem continuity (Dörr and Kollmann, 1995 ) that had not been demonstrated in previous microscopy studies or in secondary haustoria of Orobanche (Dörr and Kollmann, 1995 ). Phloem–phloem continuity has also been demonstrated in the leaf haustoria of Hyobanche (Weber, 1980 ). All haustoria thus far examined have dense parenchyma at the interface between host and parasite tissues, and these are thought to play an important role in transforming host nutrients into forms more easily utilized by the parasite. More recently, studies indicate that a macromolecular marker, green fluorescent protein, is indeed taken up from host plants by the parasites Cuscuta reflexa (Haupt, 2001 ) and Orobanche (J. Westwood, personal communication, Virginia Tech). These findings about haustorial development in other members of Orobanchaceae indicate that uptake of complex molecules such as Rubisco may be possible. Of course, the same mode of Rubisco uptake could be hypothesized for Harveya species (although this would be rather ad hoc as rbcL expression can be explained by conventional means).

This hypothesis offers an explanation to the differential deposition of the LSU observed among tissues of Harveya and Hyobanche. The movement of nutrients from host to parasite is driven by transpiration, and therefore one might expect that macromolecules obtained from the host would accumulate in the tissues with the greatest evaporative surface area. This is precisely where the LSU seems to be most concentrated in Hyobanche, in the corolla and calyx. The scale leaves of Hyobanche are quite small and probably play little role in transpiration, and the stem is nearly obsolete, or in any case, has little contact with the atmosphere. Notwithstanding, this hypothesis is yet to be tested, and haustorial characteristics of other taxa should be interpreted with caution when hypothesizing the haustorial capability of Hyobanche.

	APPENDIX

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX LITERATURE CITED

 
Taxa included in DNA sequence analyses of the rbcL operon.

Species; voucher or publication; collection source (District, Province of South Africa); GenBank accession number

Outgroups

Alectra sessiliflora; Wolfe and dePamphilis, 1998 ; —; AF026820.

Buchnera floridana; Wolfe and dePamphilis, 1998 ; —; AF026822.

Cycnium racemosum; Wolfe and dePamphilis, 1998 ; —; AF026826.^a

Cycnium racemosum; Randle; 106a (OS); Stutterheim, Eastern Cape; DQ017783.^b

Melasma scabrum; Wolfe and dePamphilis, 1998 ; —; AF190904.^a

Melasma scabrum; Randle 108 (OS); Stutterheim, Eastern Cape; DQ017784.

Harveya

H. bolusii; McMaster s.n. (OS); Gaika's Kop, Eastern Cape; DQ017773.

H. capensis; Randle 82 (NBG); Clanwilliam, Western Cape; DQ017786.

H. coccinea; Randle 144 (NBG); Cape Town, Western Cape; DQ017777.

H. huttonii; McMaster "B" (OS); Stutterheim, Eastern Cape; DQ017769.

H. hyobanchoides; Matlock s.n. (OS); Port Elizabeth, Eastern Cape; DQ017778.

H. pulchra; Edwards 1752 (OS); Underberg, Kwazulu Natal; DQ017771.

H. purpurea; Randle 79 (NBG); Caledon, Western Cape; DQ017779.

H. scarlatina; Randle 134 (NU); Underberg, Kwazulu Natal; DQ017772.

H. silvatica; Randle 136 (NU); Ngotshe, Kwazulu Natal; DQ017770.

H. speciosa; Randle 123 (OS); Elliot, Eastern Cape; DQ017774.

H. squamosa; Wolfe and dePamphilis, 1998 ; —; AF245020.^a

H. squamosa; Wolfe 710 (OS); Ysterfontein, Western Cape; DQ017775.^b

H. stenosiphon; Randle 89 (NBG); Swellendam, Western Cape; DQ017776.

Hyobanche

Hy. atropurpurea; Wolfe 716 (OS); Ceres, Western Cape; DQ017780.

Hy. glabrata; Wolfe 702 (OS); Sutherland, Northern Cape; DQ017781.

Hy. rubra; Wolfe 735 (OS); Worcester, Western Cape; DQ017787.

Hy. sanguinea; Wolfe 704 (OS); Calvinia, Northern Cape; DQ017782.

^arbcL only

^bUntranslated regions (UTRs) only

	FOOTNOTES

 
¹ The authors thank many collaborators for help in completing field work, including Jenny Archibald, Auriol Batten, Trevor Edwards, Steve Johnson, Diana and Roger Matlock, Cameron and Rhoda McMaster, Michele Pfab, Christina Potgieter, Frank Smith, Dee Snijman, and Kim Steiner; John Freudenstein, John Wenzel, and especially Mark Mort for critical comments on the manuscript; members of the Wolfe/Crawford/Freudenstein/Wenzel labs, especially Shannon Datwyler, Lisa Wallace, Jenny Archibald, Jeffery Morawetz, Shawn Krosnick, Sibyl Bucheli, and Kurt Pickett for helpful discussions; and two anonymous reviewers for critical comments. This work was supported by an International Student Dissertation Travel Grant from the Office of International Travel (OSU), the Janice Carson Beatley Herbarium Award (OSU), a Sigma Xi Graduate Research Grant, an ASPT Graduate Student Research Grant, a Graduate Student Alumni Research Award (OSU), and NSF awards DEB-0089640 and DEB-0104642.

⁴ Author for correspondence (e-mail: randle{at}ku.edu )

	LITERATURE CITED

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX LITERATURE CITED

 
Bömmer D. G. Haberhausen K. Zetsche 1993 A large deletion in the plastid DNA of the holoparasitic flowering plant Cuscuta reflexa concerning two ribosomal proteins (rpl2, rpl23), one transfer RNA (trnI) and an ORF 2280 homologue. Current Genetics 24: 171-176[CrossRef][ISI][Medline]

Bricaud C. H. P. Thalouarn S. Renaudin 1986 Ribulose 1,5-biphosphate carboxylase activity in the holoparasite Lathraea clandestina L. Journal of Plant Physiology 125: 367-370[ISI]

Bungard R. A. 2004 Photosynthetic evolution in parasitic plants: insight from the chloroplast genome. BioEssays 26: 235-247[CrossRef][ISI][Medline]

Colwell A. 1994 Genome evolution in a non-photosynthetic plant, Conopholis americana. Ph.D. dissertation, Washington University, St. Louis, Missouri, USA

Danon A. 1997 Translational regulation in the chloroplast. Plant Physiology 115: 1293-1298[CrossRef][ISI][Medline]

de la Harpe A. C. N. Grobbelaar J. H. Visser 1980 The ultrastructure of the chloroplast and the chlorophyll content of various South African parasitic flowering plants. Zeitschrift Fur Pflanzenphysiologie 100: 85-90[ISI]

de la Harpe A. C. J. H. Visser N. Grobbelaar 1981 Photosynthetic characteristics of some South African parasitic flowering plants. Zeitschrift Fur Pflanzenphysiologie 103: 265-275[ISI]

Delavault P. M. N. M. Russo N. A. Lusson P. A. Thalouarn 1996 Organization of the reduced plastid genome of Lathraea clandestina, an achlorophyllous parasitic plant. Physiologia Plantarum 96: 674-682[CrossRef]

Delavault P. V. Sakanyan P. Thalouarn 1995 Divergent evolution of two plastid genes, rbcL and atpB, in a non-photosynthetic parasitic plant. Plant Molecular Biology 29: 1071-1079[CrossRef][ISI][Medline]

dePamphilis C. W. J. D. Palmer 1990 Loss of photosynthetic and chlororespiratory genes from the plastid genome of a parasitic flowering plant. Nature 348: 337-339[CrossRef][Medline]

Dörr I. R. Kollmann 1995 Symplasmic sieve element continuity between Orobanche and its host. Botanica Acta 108: 47-55

Doyle J. J. J. L. Doyle 1987 A rapid DNA isolation procedure for small quantities of leaf tissue. Phytochemical Bulletin 19: 11-15

Ems S. C. C. W. Morden C. K. Dixon K. H. Wolfe C. W. dePamphilis J. D. Palmer 1995 Transcription, splicing, and editing of plastid RNAs in the nonphotosynthetic plant Epifagus virginiana. Plant Molecular Biology 29: 721-733[CrossRef][ISI][Medline]

Freyer R. K. Neckermann R. H. Maier H. Kossel 1995 Structural and functional analysis of plastid genomes from parasitic plants: loss of an intron within the genus Cuscuta. Current Genetics 27: 580-586[CrossRef][ISI][Medline]

García M. A. E. H. Nicholson D. L. Nickrent 2004 Extensive intra-individual variation in plastid rDNA sequences from the holoparasite Cynomorium coccineum (Cynomoriaceae). Journal of Molecular Evolution 58: 322-332[CrossRef][ISI][Medline]

Gillham N. W. J. E. Boynton C. R. Hauser 1994 Translational regulation of gene expression in chloroplasts and mitochondria. Annual Review of Genetics 28: 71-93[CrossRef][ISI][Medline]

Haberhausen G. K. Valentin K. Zetsche 1992 Organization and sequence of photosynthetic genes from the plastid genome of the holoparasitic flowering plant Cuscuta reflexa. Molecular & General Genetics 232: 154-161

Haupt S. K. J. Oparka N. Sauer S. Neumann 2001 Macromolecular trafficking between Nicotiana tabacum and the holoparasite Cuscuta reflexa. Journal of Experimental Botany 52: 173-177[Abstract/Free Full Text]

Herrmann R. G. P. Westhoff G. Link 1992 Biogenesis of plastids in higher plants. In R. G. Herrmann [ed.], Cell organelles, 274–349. Springer-Verlag, New York, New York, USA

Hibberd J. M. R. A. Bungard M. C. Press W. D. Jeschke J. D. Scholes W. P. Quick 1998 Localization of photosynthetic metabolism in the parasitic angiosperm Cuscuta reflexa. Planta 205: 506-513[CrossRef][ISI]

Kuijt J. J. H. Visser H. C. Weber 1978 Morphological observations on leaf haustoria and related organs of the South African genus Hyobanche (Scrophulariaceae). Canadian Journal of Botany 56: 2981-2986[ISI]

Kumar S. K. Tamura I. B. Jakobsen M. Nei 2001 MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 17: 1244-1245[Abstract/Free Full Text]

Kuroda H. P. Maliga 2001 Sequences downstream of the translation initiation codon are important determinants of translation efficiency in chloroplasts. Plant Physiology 125: 430-436[Abstract/Free Full Text]

Kwa S. H. Y. C. Wee P. P. Kumar 1997 Ribulose-1,5-bisphosphate carboxylase and phosphoenolpyruvate carboxylase activities of photoautotrophic callus of Platycerium coronarium (Koenig ex O. F. Muell.) Desv. under CO₂ enrichment. Plant Cell Tissue and Organ Culture 50: 75-82

Latzko E. G. J. Kelly 1983 The many-faceted function of phosphoenolpyruvate carboxylase in C-3 plants. Physiologie Vegetale 21: 805-815

Leebens-Mack J. C. dePamphilis 2002 Power analysis