An actinoporin plays a key role in water stress in the moss Physcomitrella patens

Authors


Author for correspondence:
Jeong Sheop Shin
Tel: +82 2 3290 3430
Email: jsshin@korea.ac.kr

Summary

  • • Modern land plants arose from a green algae-like ancestor c. 480 million years ago. While several novel morphological features were critical for survival in the aerial environment, physiological innovation undoubtedly played a key role in the colonization of terrestrial habitats. Recently, actinoporin genes, a small group of pore-forming toxins from sea anemones, have been found in the bryophyte and lycophyte lineages of land plants where they are upregulated in water-stressed tissues.
  • • The bryoporin gene in the moss Physcomitrella patens (PpBP) was functionally characterized by RNA blot analyses and overexpression in P. patens. In order to examine functional homology between PpBP and sea anemone actinoporins, the recombinant PpBP was subjected to hemolytic analysis of pig blood cells, which is one of the specific activities of actinoporins.
  • • PpBP was upregulated by various abiotic stresses, in particular most strongly by dehydration stress. Overexpression of the bryoporin gene heightens drought tolerance in P. patens significantly. In addition, PpBP shared the highest structural homology with actinoporins in a three-dimensional structural database and showed hemolytic activity.
  • • These results suggest that this phylogenetic distribution may have resulted from an ancient horizontal gene transfer and actinoporins may have played an important role in early land plants.

Introduction

The four main extant lineages of land plants (liverworts, mosses, hornworts, and tracheophytes) share a common green-algae-like ancestor that colonized land during the Silurian, approx. 480 million years ago (MYA) (Qiu et al., 2006). A critical precondition for the successful exploitation of terrestrial habitats was the capacity to avoid or withstand dehydration. Unlike their inferred algal ancestors, land plants possess several morphological features that prevent water loss, including a cuticle, sporopollenin, and a three-dimensional body-plan (Kenrick & Crane, 1997; Bateman et al., 1998). However, genomic data increasingly indicate that adaptation to terrestrial habitats also involved modifications of a broad array of stress response gene networks, including an increase in the number of small heat-shock proteins (Waters, 2003) and alterations of the abscisic acid (ABA)-response network (Cho et al., 2009). To date, however, the efforts to understand the evolution of physiological mechanisms to survive water-stress have largely focused on features that characterize agronomically important seed plants (Wang et al., 2008; Choudhary et al., 2009; Urano et al., 2009). Whether the components of seed-plant stress response networks are modern representatives of the proteins or pathways that initially allowed plants to survive on land, or are subsequent elaborations that evolved much later, is largely unknown (but see Marella et al., 2006).

The availability of genomic databases for several lineages of land plants has facilitated powerful transcriptomic or proteomic screens enabling investigators to identify large numbers of loci that are differentially regulated during drought stress. One particularly surprising outcome of several such studies was the discovery of actinoporin-like proteins (ALPs) in drought stress expressed sequence tag (EST) libraries from the mosses Tortula ruralis (Oliver et al., 2004), Physcomitrella patens (Nishiyama et al., 2003) and the lycophyte Selaginella lepidaphylla (Iturriaga et al., 2006). Actinoporins are a small but well-studied group of highly cytotoxic proteins. The most extensively characterized representative, equinatoxin II (EqtII), is a pore-forming toxin isolated from nematocysts (stinging cells) from the sea anemone Actinaria equina. EqtII binds with high specificity to sphingomyelin within the exoplasmic face of the cell membrane (Anderluh & Macek, 2002) and oligomerizes to form a tetrameric pore (Malovrh et al., 2000; Mally et al., 2002). In mammals and crustaceans EqtII elicits erythrocyte hemolysis (Anderluh et al., 1996; Malovrh et al., 2000; Wang et al., 2000; Anderluh & Macek, 2002; Nagai et al., 2002), platelet aggregation (Teng et al., 1988), pulmonary edema (Lafranconi et al., 1984) and cardiac arrest (Bunc et al., 1999).

Until recently, actinoporins were thought to be restricted to sea anemones (Cnidaria: Actiniidae). However, ALPs now have been identified in other cnidarians and a wide variety of vertebrates, in addition to a small number of land plants, and may be homologous to fungal fruit body lectins (Rensing et al., 2005; Gutierrez-Aguirre et al., 2006) [Correction added after online publication 10th September 2009: in the preceding sentence, the reference citation ‘Lang et al., 2005’ was corrected to ‘Rensing et al., 2005’.]. Here, we show that a bryophyte actinoporin protein (bryoporin) in P. patens, PpBP, is upregulated under water stress conditions, and that overexpressing this protein heightens drought resistance in this species. Since bryoporin is found in EST libraries generated under similar conditions in the lycophyte S. lepidaphylla, as well as another moss, the evidence suggests that the bryoporin played an important role in water stress in the common ancestor of these plants, one of the earliest land plants. We also provide structural and functional evidence to suggest that bryoporins and true actinoporins are homologous, and discuss scenarios to explain the unusual phylogenetic distribution of this protein.

Materials and Methods

Plant material and sample treatments

Physcomitrella patens was cultured in the Knop's liquid medium supplemented with ammonium tartrate (Knop's-AT) (Ashton & Cove, 1977) by shaking at 125 rpm for 15 d; gametophores were developed in Knop's-AT solid media for 1 month. Both were placed at 25 ± 0.5°C under continuous light (20 W m−2) (Ashton & Cove, 1977). The cultured tissues were collected using a sieve (200 mesh; Sigma), frozen in liquid nitrogen, and stored at −70°C for the experiments.

For the experiments involving various abiotic stresses, 15-d-old protonemal and gametophore tissues were transferred to fresh Knop's liquid media supplemented with 100 µm ABA, 300 mm NaCl, 600 mm mannitol, 50 µm H2O2, 100 µm jasmonic acid (JA), 100 µm ethylene (ET), 100 µm HgCl2, 100 µm CuSO4, 100 µm salicylic acid (SA), 100 µm gibberellic acid (GA), 10 µm indole-3-acetic acid (IAA) or 10 µm 6-benzyladenine (BA), respectively, and cultured with shaking under the same conditions as described in the preceding paragraph. Cold treatments were performed by culturing 15-d-old protonemata at 4°C with shaking at 125 rpm under continuous light and exposure of gametophore plants grown at 22°C to a temperature of 4°C. The dehydration treatment was performed by harvesting the plants from the culture, removing attached medium and drying the plants on 3M filter (Whatman) paper at 25°C. For the wounding treatment, protonemal and gametophore tissues were blended and scratched by forceps, respectively.

DNA sequence analysis

A full-length cDNA clone (National Center for Biotechnology Information (NCBI) Accession number BJ176008) coding for bryoporin was kindly provided by Professor Tadasu Shin-I (National Institute of Genetics, Mishima, Japan). Multiple sequence alignment was performed by the Clustal V method (Higgins & Sharp, 1989) using the megalign program (version 4.0; DNASTAR, Madison, WI, USA). The sequences and their accession numbers in NCBI are as follows: Physcomitrella patens 1a, AAV65396; Physcomitrella patens 1b, XP_001772366; Physcomitrella patens 2, XP_001762857; Astatotilapia burtoni, DY630341; Actinia equina (II), AAC47005; Actinia equina (IV), AAD39836; Actinia equina (V), AAC05720; Actinia tenebrosa, P61915; Actinia. villosa, BAD74019; Danio. rerio, EH473956; Heteractis crispa (I), AAW47930; Heteractis crispa (II), P0C1F8; Heteractis magnifica, AF170706; Oryzias. latipes, BJ724629; Oncorhynchus. mykiss, CX151000; Oulactis orientalis, AAW47579; Paralabidochromis chilotes, BJ676078; Phyllodiscus semoni, BAC45007; Stichodactyla helianthus (I), CAC00651; Stichodactyla helianthus (II), CAC20912; Selaginella lepidophylla, BM402569; Selaginella moellendorffii, CAOS6349; Sagartia rosea, AAP04347; Salmo salar, EG832538; Syntrichia ruralis, CN209292.

RNA gel blot analysis

Total RNA was isolated from protonemal and gametophore samples with TRIZOL reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. Fifteen micrograms of total RNA was fractionated in denaturing formaldehyde gel and blotted onto Tropilon-plus nylon membrane (Tropix, Foster City, CA, USA). The RNA blots were hybridized with the dCTP-biotin-labeled full-length cDNA probes for 16 h at 67°C. The membranes were washed twice with 2× standard saline citrate (SSC)/1% sodium dodecyl sulfate (SDS) at room temperature for 5 min each, 0.1× SSC/1% SDS at 67°C for 15 min and 1× SSC at room temperature for 5 min. The hybridized probes were detected using the Southern-Star (Applied Biosystems, Tropix, CA, USA) according to the manufacturer's instructions. Membranes were exposed to a LAS3000 (Fujifilm, Tokyo, Japan).

Expression of recombinant bryoporin protein in Escherichia coli

The coding region of PpBP cDNA was amplified by PCR with a set of primers (5′-CATGCCATGGGGGATGGCCGAAGCAATTATT-3′/5′-CGGGATCCTAAGCCTTCTTC AGTTCCAC-3′). Amplification was performed for 30 cycles at 94°C for 1 min, 55°C for 1 min and 72°C for 1 min, followed by 72°C for 7 min. The PCR product was digested with NcoI and BamHI, and cloned into the pET-32c(+) vector (Novagen Inc., Madison, WI, USA). The recombinant PpBP protein was produced in E. coli strains BL21 (DE3) pLysS, and then this His-tagged protein was purified using Ni-NTA chromatography (Qiagen, Valencia, CA, USA) following the manufacturer's manual.

Preparation of bryoporin polyclonal antibody

The fusion recombinant protein of PpBP was digested with the enterokinase (Sigma, St. Louis, MO, USA) and separated by SDS-polyacrylamide gel electrophoresis (PAGE). The electroeluted PpBP recombinant protein was used as the immunogen to generate mouse antiserum. Six to eight-week-old mice (approx. 18–20 g) were immunized intraperitoneally with 100 µg antigen emulsified in 2.5 mg of aluminum hydroxide gel (Pierce, Rockford, IL, USA) on the day 1, followed by a second immunization on the day 11. Seven days after immunization, blood was collected from mice hearts and centrifuged. The supernatant (antiserum) was collected and used as a bryoporin polyclonal antibody.

Enzyme-linked immuno-sorbent assay (ELISA)

An ELISA was performed with serial antigen dilution (0, 5, 10, 20, 40, 80, 160, 320 ng ml−1) of the PpBP recombinant protein, and protonemal and gametophore crude extracts. Proteins mixed with coating buffer (10 mm Na2CO3 and 41 mm NaHCO3) were coated onto the plate for 2 h at 37°C. After washing with phosphate-buffered saline (PBS) (11 mm Na2HPO4, 1.3 mm NaH2PO4, 0.15 m NaCl) containing 0.05% Tween-20 followed by blocking with PBS containing 1% BSA, the mouse anti-PpBP antibody (1/500 dilution) was added and incubated for 1 h at 37°C. After washing with PBS containing 0.05% Tween-20, the horseradish peroxidase (HRP)-conjugated anti-mouse IgG (1/5000 dilution) was added and incubated for 1 h at 37°C, followed by the addition of tetramethylbenzidine as a substrate. After adding stop solution (1 m H3PO4), the absorbance was measured at A450.

Hemolytic assay

Fresh pig blood was collected into a solution of 10% EDTA and centrifuged at 3000 g for 5 min at room temperature. In order to remove the plasma and buffy coat, the cells were treated with PBS repeatedly until the supernatant become clear. The PpBP recombinant proteins and total proteins from protonemal and gametophore tissues were prepared in the concentration of 50, 100, 200 and 400 µg µl−1, and then mixed with 0.5% (v : v) erythrocyte suspension in PBS, followed by incubation for 4 h at 25°C. After centrifugation at 3000 g for 10 min, the concentration of hemoglobin released into the supernatant was measured on a microplate reader at A540. Saponin (Sigma, St. Louis, MO, USA) was used as a positive control to compare the hemolytic activity by the addition of final concentration of 100 µg ml−1 into 0.5% (v : v) erythrocyte suspension.

The inhibitory effect of sphingomyelin and other lipids on hemolytic activity was estimated by measuring the residual hemolytic activity of unbound PpBP. The PpBP protein (400 ng µl−1) was preincubated for 10 min at room temperature with sphingomyelin, cholesterol, 1,2-dipalmitoyl-sn-glycero-3-phosphocholine, or 1,2-dioleoyl-sn-glycero-3-phosphocholine, respectively, at a final concentration of 33 µg ml−1 (Anderluh & Macek, 2002; Sakurai et al., 2004) and used for hemolytic assay as already described.

Vector construction and plant transformation

The pLUG3 vector was kindly provided by Dr David Cove (Leeds University, Leeds, UK), which contained the hph gene driven by the constitutive CaMV 35S promoter and the NOS terminator. The CaMV35S:PpBP plasmid was constructed by amplifying the entire coding region of PpBP by PCR using a set of primers (5′-GCTCTAGAGCGATGGCCGAAGCA ATTATT-3′/5′-GCTCTAGAGCACTAAGCCTTCTTCAGTTCCAC-3′), and cloned into a XbaI-digested pLUG3 vector in the sense orientation. This construct was transformed into the moss as described in Cho et al. (1999).

Measurement of chlorophyll contents

Samples were ground in liquid nitrogen and vigorously mixed with 80% (v : v) acetone. In order to exclude cell debris, the samples were centrifuged for 5 min at 11000 g. The supernatant was subjected to measurement of chlorophyll contents at A663 and A645. After measurement, the supernatant was mixed with cell debris and dried completely in a speed-vac to obtain dry weight. The total chlorophyll content was calculated using the equation: µg chlorophyll g−1 DW = ((A663)(0.00802) + (A645)(0.0202)) × 1.5/DW.

Results and Discussion

Bryoporin is induced by water stress

Actinoporins recently have been reported in drought stress EST libraries in S. lepidaphylla (Iturriaga et al., 2006), T. ruralis (Oliver et al., 2004) and P. patens (Nishiyama et al., 2003). To more rigorously characterize the function of a plant actinoporin (bryoporin), we surveyed the expression pattern of bryoporin in the model moss P. patens (PpBP). The haploid gametophyte stage of the P. patens life cycle consists of filamentous protonemata (which germinate from a spore) and erect leafy gametophores that develop on the protonemata. Northern blot analysis showed that very little PpBP mRNA was found in 15-d-old protonemata, but the transcript was present in young gametophores and reached a maximal level in mature gametophores (Fig. 1a). An ELISA with a mouse antiserum against PpBP demonstrated that the abundance of the protein throughout the development of P. patens was proportional to the production of the transcript (see the Supporting Information, Fig. S1a).

Figure 1.

Bryoporin is induced by water stress. RNA gel blot analyses on PpBP were performed under various conditions. Ribosomal RNAs stained with ethidium bromide were used as an RNA loading control. (a) Specific expression of PpBP in gametophores. P15, 15-d-old protonemal tissue; G20, G25 and G30, represent 20-, 25- and 30-d-old gametophores, respectively; Gm, fully grown mature gametophore. (b) Induction of PpBP in protonemata under various abiotic stresses. Protonemata were treated with abscisic acid (ABA), NaCl, mannitol, H2O2, jasmonic acid (JA), ethylene (ET), HgCl2, CuSO4, salicylic acid (SA), gibberellic acid (GA), indole-3-acetic acid (IAA), or 6-benzyladenine (BA), respectively. Cold treatments were performed by culturing protonemata at 4°C. Mock, control with no treatment. (c) Induction of PpBP in gametophore tissues by water stress. (d) Suppression of PpBP induction by liquid culture. Gametophores grown on solid media were transferred to liquid culture. Samples were harvested after the designated time.

The presence of bryoporin in plant desiccation EST libraries (Oliver et al., 2004; Iturriaga et al., 2006) suggested that the protein may be a part of a stress response pathway. Unlike the protonemal tissue, which grows in close contact with its generally humid substrate, the leafy gametophores where PpBP is natively expressed are exposed to the aerial environment. To determine whether PpBP transcription was induced by environmental stress (Cho et al., 2007), we surveyed PpBP expression patterns in protonemal tissues that were subjected to various stresses and phytohormones (Fig. 1b). The transcript was strongly induced by growth on media containing the plant hormone ABA or mannitol, and by dehydration, but not salt or cold. Similarly, mechanical wounding of plant tissue or treatment with the plant hormones JA or SA also induced transcription, while other treatments had no significant effect on PpBP transcript levels. Dehydration and treatment with exogenous ABA or mannitol rapidly induced PpBP expression in gametophores; dehydration caused the most rapid change in transcription, raising PpBP to its maximal level at 12 h (Fig. 1c).

To further test the association between desiccation stress and bryoporin, we induced PpBP expression by a drought stress, and then transferred the stressed plants to a liquid medium. The PpBP mRNA declined rapidly, and was absent 12 h after transfer (Fig. 1d). Finally, we subjected wild type and PpBP over-expression lines to periods of desiccation, followed by recovery, and then measured their chlorophyll content as an indicator of viability. The overexpression of PpBP resulted in a two-fold to five-fold increase in chlorophyll content (Fig. 2a) and a general increase in viability (Fig. 2b). Although the quantity of PpBP mRNA was increased only slightly in the transgenic lines (Fig. S2a), the crude gametophore extract showed an approximately five-fold increase in protein activity over that of wild-type plants (Fig. S2b). These results, combined with the inferred presence of bryoporin in the ancestor of bryophytes and lycophytes, suggests that physiological mechanisms of stress tolerance, in addition to morphological adaptations, may have been critical for the colonization of terrestrial habitats by early land plants (Waters, 2003). Our attempts to silence gene expression using an RNAi strategy were unsuccessful, but we are currently creating knockout lines using gene targeting in order to identify its interacting partners and mechanism of action (Quatrano et al., 2007).

Figure 2.

Bryoporin overexpression confers dehydration tolerance to Physcomitrella patens. Three individual PpBP overexpression lines driven by CaMV 35S promoter were harvested, and dehydrated for 1, 2 or 3 d, rehydrated for 1 h and then transferred to the normal growth condition. (a) Chlorophyll contents of dehydrated mosses. Chlorophyll contents were measured after 15 d of recovery. Error bars are standard deviation of two replications, and each replication comprises three sub-replications. (b) Dehydration tolerance of transgenic mosses. Photos were taken after 2 wk of recovery. Bars, 1 cm.

PpBP is an actinoporin protein

The land plant bryoporins form a well-supported clade with the sea anemone actinoporins, exclusive of other cnidarian or metazoan ALPs (Figs 3, S3) (Gutierrez-Aguirre et al., 2006), an arrangement that might have arisen by convergent evolution. However, PpBP shares 50% amino acid identity with actinoporin EqtII and a conserved sphingomyelin-binding domain not found in other sequenced plant genomes. Indeed, when the PpBP amino acid sequence is threaded through the crystal structure of EqtII (Hinds et al., 2002) or that of a related actinoporin (Mancheno et al., 2003) it coincides perfectly (http://toolkit.tuebingen.mpg.de/hhpred), suggesting that the plant and cnidarian proteins may have similar biochemical properties.

Figure 3.

 Evolutionary relationships among actinoporin-like protein (ALPs). (a) Multiple sequence alignments of deduced amino acids of bryoporin with previously reported actinoporin proteins. Residues were annotated as identical (black), 80% similar (dark gray), 60% similar (light gray), or dissimilar (no color). The dashed (–) regions indicated the gap residues. The red and the green boxes at the C-terminal end indicate the Trp cluster for sphingomyelin binding (Bakrac et al., 2008) and Trp residue for efficient hemolytic activity (Malovrh et al., 2000), respectively. (b) Phylogenetic relationships among the main lineages of ALPs, with maximum likelihood bootstrap values indicated (adapted from Gutierrez-Aguirre et al., 2006).

To test whether this physical similarity reflects conserved functionality, we assayed the hemolytic activity of purified recombinant PpBP protein (shown in Fig. S4), and crude protonemal or gametophore extracts (Figs 4a and S1b). Both the recombinant PpBP protein and the total gametophore extract containing native bryoporin caused 50% hemolytic activity at concentrations of 767.8 ng µl−1 and 640.8 ng µl−1, respectively. No hemolytic activity was present in the crude extract of protonemata, indicating that hemolysis was not a result of a constitutively expressed moss protein. The activity of both the recombinant PpBP and the gametophore extract was inhibited by incubation with sphingomyelin (but not other membrane components) (Fig. 4a,b). Thus, unlike the fish actinoporin Dr1, the P. patens bryoporin has a binding specificity for sphingomyelin that is similar to that of sea anemone actinoporins (Gutierrez-Aguirre et al., 2006). The activity in the gametophore extract, however, was 1.3–400 times lower than that of the sea anemone actinoporins (Wang et al., 2000; Nagai et al., 2002), potentially because PpBP lacks two critical tryptophan residues corresponding to Trp-45 and Trp-149 of EqtII for efficient hemolytic activity (Figs 3a, S3) (Malovrh et al., 2000; Wang et al., 2000; Bakrac et al., 2008). We suspect that PpBP is adapted to bind a phosphocholine derivative with the unique amido or hydroxyl groups found in sphingomyelin (to which actinoporins also specifically bind) (Bakrac et al., 2008), as plants lack sphingomyelin (Lynch & Dunn, 2004).

Figure 4.

 Hemolytic activity of bryoporin is specifically inhibited by sphingomyelin, the sea anemone actinoporin specific binding phospholipid. Recombinant PpBP protein (rBP) and crude extracts from protonemata (PCE) and gametophores (GCE) were incubated with pig erythrocytes. The concentration of hemoglobin released into the supernatant was measured at 540 nm. Data are represented as the means. Error bars are standard deviation of two replications, and each replication comprises three subreplications. (a) Inhibition of hemolytic activities by sphingomyelin. Protein samples were incubated with sphingomyelin, and then hemolytic activity was determined (open bars). Closed bars represent the natural hemolytic activity of each sample. (b) Specificity of sphingomyelin inhibition of bryoporin hemolytic activity. rBP was preincubated with a various phospholipids, sphingomyelin, cholesterol, 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPGP), or 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOGP), respectively. Hemolytic activities were then measured.

The evolution of plant actinoporins

The amino acid sequence, structural, and functional data suggest that PpBP and EqtII are homologs. While it is possible that the common ancestor of land plants and sea anemones contained a true actinoporin gene, for this hypothesis to be true actinoporin homologs subsequently would have been lost in most eukaryotic lineages (e.g. angiosperms, green algae, protists, fungi, mammals and insects) where ALPs are absent from the completed genome sequences. The conservation of such a specialized protein in only the sea anemones and a few lineages of land plants, however, would be highly enigmatic.

Alternatively, the unusual phylogeny of ALPs could result from the horizontal transfer of an ancestral actinoporin between a metazoan and a green plant. Horizontal gene transfer (HGT) frequently contributes to adaptive evolution in prokaryotes (Boucher et al., 2003; Gogarten & Townsend, 2005), and its influence on eukaryotic genome content is increasingly recognized (Andersson et al., 2003; Keeling & Inagaki, 2004; Steele et al., 2004; Diao et al., 2006; Richards et al., 2006; Richardson & Palmer, 2007; Emiliani et al., 2009). To explain the sister relationship between land plants and sea anemone ALPs, an HGT event would have had to occur in the Ordivician, after the divergence between unicellular green algae and land plants, but before the divergence of bryophytes and tracheophytes. Fossils assignable to modern cnidarian orders were present by the Cambrian, and many cnidarian species maintain symbiotic relationships with photosynthetic algae, potentially providing the opportunity for lateral transfer of an ancestral actinoporin. Interestingly, the fungal fruit body lectins, the putative sister group to the ALPs, contain closely related functional genes from a moss and a liverwort (Oliver et al., 2004; Peumans et al., 2007), suggesting that genes with unusual bioactivities may be particularly likely to have similarly unusual phylogenetic distributions (Dobrindt et al., 2004; Cordes & Binford, 2006; Richards et al., 2006; Sanders, 2006).

Acknowledgements

We thank Heung Kyu Kim (Korea University, Seoul Korea) and Hong Je Park (Komed Co., Kyunggi, Korea) for their help in preparing the bryoporin antibody and performing the ELISA, and Dr Robert Blankenship (Washington University, USA) and Dr Michael Axtell (Pennsylvania State University, USA) for their critical reading and valuable comments on our manuscript. This work was supported by a grant from the BioGreen21 Program of the Rural Development Administration (to J.S.S.), by a grant from the Crop Functional Genomics Center of the 21st Century Frontier Research Program (to J.S.S.), by the Korea Research Foundation Grant (MOEHRD, Basic Research Promotion Fund M01-2004-000-10317-0 to S.H.C.), by Washington University (to R.S.Q.) and by a National Institutes of Health National Research Service Award (5 F32 GM075606-02 to S.F.M.).

Ancillary