Tag Archives: Robert E. Marc

Rod Bipolar Cell Networks in Early Retinal Remodeling

Rebecca Pfeiffer, a post-doc in the laboratory presented her work on “Rod Bipolar Cell Networks in Early Retinal Remodeling” as a platform presentation at the ISER 2018 meeting in Belfast, Northern Ireland.

Authors: Rebecca Pfeiffer, James R. Anderson, Daniel P. Emrich, Jeebika Dahal, Crystal L Sigulinsky, Hope AB Morrison, Jia-Hui Yang, Carl B. Watt, Kevin D. Rapp, Jessica C Garcia, Mineo Kondo, Hiroko Terasaki, Robert E. Marc, and Bryan W. Jones.

Abstract: Retinal remodeling is a form of negative plasticity that occurs as a consequence of retinal degenerative diseases. Part of retinal remodeling involves anomalous sprouting of processes, termed neurites. The synaptic structures and partners of the neurites are not yet defined, leading to uncertainty about the consistency of network motifs between healthy and degenerate retina. Our goal is to map out the identities and network relationships of bipolar cell networks using a connectomics strategy. Retinal connectomes or ultrastructural maps of neuronal connectivity have substantially contributed to our understanding of retinal network topology, providing ground truth against which pathological network topologies can be evaluated. We have generated the first pathoconnectome (RPC1), or connectome of pathological tissues, of early retinal remodeling at 2nm/pixel, and are currently investigating the impact of remodeling on network architecture.
The tissue for RPC1 was obtained from a 10mo transgenic P347L rabbit model of autosomal dominant retinitis pigmentosa. Tissue was fixed in mixed aldehydes, osmicated, dehydrated, embedded in epon resin, and sectioned at 70nm. Serial sections were placed on grids, stained, and imaged using a JEOL JEM-1400 TEM using SerialEM software. Every 30th section was reserved for computational molecular phenotyping (CMP), and probed for small molecules: glutamate, glutamine, glycine, GABA, taurine, glutathione; or TEM compatible proteins GFAP and GS. The pathoconnectome volume is explored and annotated using the Viking software suite.
RPC1 was selected as an example of early retinal remodeling, demonstrating Muller cell hypertrophy, metabolic dysregulation, and degeneration of rod outer segments, indicating phase 1 remodeling and neuronal sprouting. We have observed the presence of both cone pedicles and rod spherules within the OPL to be synaptically active with neurites from some rod bipolar cells forming functional synapses with both rod spherules and cone pedicles. These rod bipolar cells also exhibit structurally altered ribbon synapses. We are currently evaluating network motifs and comparing them to networks established from our previous connectome, RC1, generated from a healthy rabbit.
These findings allow us to evaluate and analyze the impact of retinal remodeling on retinal networks which may have important implications for therapeutic interventions being developed which rely on inner retina network integrity.

Pathoconnectome Analysis of Müller Cells in Early Retinal Remodeling

Rebecca Pfeiffer, a post-doc in the laboratory presented her work on “Pathoconnectome Analysis of Müller Cells in Early Retinal Remodeling” as a platform presentation at the RD2018 meeting in Killarney, Ireland.

Authors: Rebecca Pfeiffer, James R. Anderson, Daniel P. Emrich, Jeebika Dahal, Crystal L Sigulinsky, Hope AB Morrison, Jia-Hui Yang, Carl B. Watt, Kevin D. Rapp, Mineo Kondo, Hiroko Terasaki, Jessica C Garcia, Robert E. Marc, and Bryan W. Jones.

Purpose: Glia play important roles in neural system function. These roles include, but are not limited to: amino acid recycling, ion homeostasis, glucose transport, and removal of waste. During retinal degeneration, Muller cells, the primary macroglia of the retina, are one of the first cells to show metabolic and morphological alterations in response to retinal stress. The metabolic alterations observed in Muller cells appear to manifest in regions of photoreceptor degeneration; however, the precise mechanisms that govern these alterations in response to neuronal stress, synapse maintenance, or glia-glia interactions is currently unknown.  This project aims to reconstruct Muller cells from a pathoconnectome of early retinal remodeling at 2nm/pixel with ultrastructural metabolic data to determine the relationship of structural and metabolic phenotypes between neighboring neurons and glia.

Methods:  Retinal pathoconnectome 1 (RPC1) is the first connectome to be assembled from pathologic neural tissue (a pathoconnectome). The tissue selected for RPC1 was collected post mortem from a 10 month transgenic P347L rabbit model of autosomal dominant retinitis pigmentosa, fixed in 1% formaldehyde, 2.5% glutaraldehyde, 3% sucrose, and 1mM MgSO4 in cacodylate buffer (pH 7.4). The tissue was subsequently osmicated, dehydrated, resin embedded, and sectioned at 70nm. Sections were placed on formvar grids, stained, and imaged at 2nm/pixel on a JEOL JEM-1400 TEM using SerialEM software. 1 section was reserved from every 30 sections for CMP, where it was placed on a slide and probed for small molecules: glutamate, glutamine, glycine, GABA, taurine, glutathione; or TEM compatible proteins GFAP and GS. The pathoconnectome volume was evaluated and annotated using the Viking software suite.

Results: RPC1 demonstrates hallmarks of early retinal degeneration and remodeling, including the glial phenotypes of hypertrophy and metabolic variation between neighboring Muller cells. Early evaluation of these glia demonstrates variations in osmication in Muller cells as well as apparent encroachment of glial end-feet on one another.  We are currently in the process of reconstructing multiple Muller cells within RPC1 and their neighboring neurons.  Once complete, we will assess the relationship between Muller cell phenotype and the phenotypes of contacted neuronal and glial neighbors.

Conclusions: How neural-glial relationships are affected by retinal remodeling may help us understand why remodeling and neurodegeneration follow photoreceptor degeneration. In addition, determining these relationships during remodeling will be crucial to developing therapeutics with long-term success. RPC1 provides a framework to analyze these relationships in early retinal remodeling through ultrastructural reconstructions of all neurons and glia in an intact retina. These reconstructions, informed by quantitative metabolite labeling, will allow us to evaluate these neural-glial interactions more comprehensively than other techniques have previously allowed.

Coupling Architecture Of The Aii/ON Cone Bipolar Cell Network In The Degenerate Retina

Crystal Sigulinsky, a post-doc in the lab, presented her work on “coupling architecture of the
Aii/ON cone bipolar cell network in the degenerate retina” at the RD2018 meeting in Killarney, Ireland today.  Authors are: Crystal L Sigulinsky, Rebecca L Pfeiffer, James R Anderson, Jeebika Dahal, Hope Morrison, Daniel P. Emrich, Jessica C Garcia, Jia-Hui Yang, Carl B. Watt, Kevin D. Rapp, Mineo Kondo, Hiroko Terasaki, Robert E. Marc, and Bryan W. Jones.

Purpose: Retinal network hyperactivity within degenerative retinal networks is a component of the disease process with implications for therapeutic interventions for blinding diseases that depend upon the surviving retinal network. Connexin36-containing gap junctions centered on the Aii amacrine cell network appear to mediate the aberrant signaling observed in mouse models of retinal degeneration. However, it remains unclear whether this hyperactivity reflects changes in the underlying circuitry or dysfunction/dysregulation of the normative circuitry. Mapping retinal circuitry in the ultrastructural rabbit Retinal Connectome, RC1, has revealed Aii network topologies explicitly involving gap junctions. In addition to canonical Aii-to-Aii and Aii-to-ON cone bipolar cell (CBC) coupling, we describe pervasive in- and cross-class coupling motifs among ON CBCs that extend and dramatically expand the coupled Aii network topologies. Since virtually every gap junction in the inner plexiform layer contains Connexin36, these circuits likely participate in the aberrant signaling of degenerate retinas. This study examines these Aii and ON CBC coupling motifs in Retinal PathoConnectome 1 (RPC1), an ultrastructural pathoconnectome of a rabbit model of retinitis pigmentosa.

Approach: RPC1 is a 2nm/pixel resolution volume of retina from a 10 month old, transgenic P347L rabbit model of autosomal dominant retinitis pigmentosa in early phase 1 retinal remodeling, a time point where cone and rod photoreceptors are still present, albeit going through cell stress. RPC1 spans the vitreous to basal outer nuclear layer and was built by automated transmission electron microscopy and computational assembly. ON CBCs, Aii amacrine cells, and their coupling partners were annotated using the Viking application and explored with 3D rendering and graph visualization of connectivity. Gap junctions were validated by 0.25 nm resolution recapture with goniometric tilt when necessary. Motifs were compared to those discovered in RC1. RC1 is a 2 nm resolution, 0.25 mm diameter volume of a light-adapted adult female Dutch Belted rabbit retina spanning the ganglion cell through inner nuclear layers.

Conclusions: RPC1 shows degeneration of rod outer segments, Müller cell hypertrophy and neuronal sprouting, characteristic of early stage retinal degeneration and phase 1 remodeling, when retinal hyperactivity and its reliance on gap junctional coupling has likely already initiated and human patients would still have some vision. All major coupling motifs (Aii-to-Aii, Aii-to-ON CBC, and ON CBC-to-ON CBC) were observed. Preliminary examinations indicate that several ON CBC classes retained their class-specific coupling profiles, accepting and rejecting specific combinations of Aii and ON CBC class partnerships. However, recent findings reveal aberrant partnerships in the coupled network, including both loss of prominent motifs and acquisition of novel ones. Thus, clear aberrant morphological and synaptic changes have been identified in RPC1, including changes in the coupling specificity and gap junction distributions of both Aii amacrine cells and ON CBCs (Figure 6). This suggests that the Aii/ON CBC circuit topology is already altered during early phase 1 remodeling, with substantial implications for therapeutic interventions in human subjects. The full coupling network is actively being examined and progress has begun on RPC2, a second pathoconnectome for examining later, phase 2 remodeling in this same model.

An almost full size poster available here in pdf format.

A Pathoconnectome of Early Retinal Remodeling

This abstract was presented today, Monday, April 30th at the 2018 Association for Research in Vision and Opthalmology (ARVO) meetings in Honolulu, Hawaii by Rebecca Pfeiffer, Robert E. Marc, James R. Anderson, Daniel P. Emrich, Carl B. Watt, Jia-Hui Yang, Kevin D. Rapp, Jeebika Dahal, Mineo Kondo, Hiroko Terasaki, and Bryan W. Jones.

Retinal remodeling is a consequence of retinal degenerative disease, during which neurons sprout new neurites whose synaptic structures and partners are not yet defined. Simultaneously during remodeling, Müller cells (MCs) undergo structural and metabolic changes, whose impact on surrounding neurons is an active area of research. Retinal connectomes have elucidated and validated fundamental networks. These data provide further classification of neuronal types and subtypes and a precise framework for modeling of retinal function, based on ground truth networks. The creation of the first pathoconnectome (RPC1), a connectome from pathological retinal tissue, provides the opportunity to determine connectivites between neurons, while simultaneously evaluating glial remodeling. Computational Molecular Phenotyping (CMP) embedded within the ultrastructure provides metabolic factors of pathologies.

RPC1 was collected post-mortem from a 10mo TgP347L rabbit model of adRP, fixed in 1% FA, 2.5% GA, 3% sucrose, and 1mM MgSO4 in cacodylate buffer (pH 7.4). The tissue was osmicated, dehydrated, resin embedded, and sectioned at 70nm. Sections were placed on formvar grids, stained, and imaged on a JEOL JEM-1400 TEM using SerialEM. 1 section was reserved from every 30 section for CMP, where it was probed for small molecules: glutamate, glutamine, glycine, GABA, taurine, glutathione; or proteins GFAP and GS. RPC1 was evaluated using the Viking software suite.

RPC1 was chosen based on early features of retinal degeneration/remodeling: degeneration of rod OS, MC hypertrophy, and neuronal sprouting. RPC1 consists of 948 serial sections spanning the ONL to the vitreous, with a diameter of 90µm. We find dendrites extending from rod bipolar cells to cone pedicles, originally described in light microscopy, and active synaptic contacts. We also see alterations of synaptic structure in the IPL, and MC morphological changes affecting surface to volume and neuron/glial relationships. Network motifs are being actively investigated.

We observe many features of remodeling previously described using light microscopy, and confirm active synaptic contact. We also find synaptic structural features, not previously described. In addition, early evaluation of MC morphology demonstrates marked changes in MC shape and associations with nearby neurons and glia, which, combined with CMP, will be instrumental in understanding how MCs affect retinal remodeling.

Corticostriatal Circuit Defects in Hoxb8 Mutant Mice

We have a new publication in Molecular Psychiatry, Corticostriatal circuit defects in Hoxb8 mutant mice. (Direct link here).  Authors are:  Naveen Nagarajan, Bryan W. Jones, Peter West, Robert Marc, and Mario R. Capecchi.

Abstract: Hoxb8 mutant mice exhibit compulsive grooming and hair removal dysfunction similar to humans with the obsessive-compulsive disorder (OCD)-spectrum disorder, trichotillomania. As, in the mouse brain, the only detectable cells that label with Hoxb8 cell lineage appear to be microglia, we suggested that defective microglia cause the neuropsychiatric disorder. Does the Hoxb8 mutation in microglia lead to neural circuit dysfunctions? We demonstrate that Hoxb8 mutants contain corticostriatal circuit defects. Golgi staining, ultra-structural and electrophysiological studies of mutants reveal excess dendritic spines, pre- and postsynaptic structural defects, long-term potentiation and miniature postsynaptic current defects. Hoxb8 mutants also exhibit hyperanxiety and social behavioral deficits similar to mice with neuronal mutations in Sapap3, Slitrk5 and Shank3, reported models of OCD and autism spectrum disorders (ASDs). Long-term treatment of Hoxb8 mutants with fluoxetine, a serotonin reuptake inhibitor, reduces excessive grooming, hyperanxiety and social behavioral impairments. These studies provide linkage between the neuronal defects induced by defective Hoxb8-microglia and neuronal dysfunctions directly generated by mutations in synaptic components that result in mice, which display similar pathological grooming, hyperanxiety and social impairment deficits. Our results shed light on Hoxb8 microglia-driven circuit-specific defects and therapeutic approaches that will become essential to developing novel
therapies for neuropsychiatric diseases such as OCD and ASDs with Hoxb8-microglia being the central target.

Synaptic Inputs To A Gamma Ganglion Cell In Rabbit Retina

This abstract was presented today, May 8th at the 2017 Association for Research in Vision and Opthalmology (ARVO) meetings in Baltimore, Maryland by Andrea Bordt, Diego Perez, Robert E. Marc, James R. Anderson, Carl B. Watt, Bryan W. Jones, Crystal Sigulinsky, James S. Lauritzen, Danny Emrich, Noah Nelson, Luke Tseng, Weiley Liu, and David W. Marshak. Full resolution version here.

Purpose: There are at least 30 distinct types of mammalian retinal ganglion cells, each sensitive to different features of the visual environment, and these can be grouped according to their morphology. One such group, the gamma cells, was identified more than 40 years ago, but their synaptic inputs have never been described. That was the goal of this study.

Methods: The synaptic inputs to a subtype of gamma cell with dendrites ramifying in the outer sublamina of the inner plexiform layer (IPL) of the rabbit retina were identified in a retinal connectome developed using automated transmission electron microscopy.

Results: The gamma cell was always postsynaptic in the IPL, confirming its identity as a ganglion cell. The local synaptic input should produce relatively weak OFF reposnses to stimuli confined to the center of the gamma cell’s receptive field. It typically received only one synapse per bipolar cell from at least 4 types of OFF bipolar cells. Because bipolar cells vary in their response kinetics and contrast sensitivity. each type would provide a small, asynchronous excitatory input. The amacrine cells at the dyad synapses provided only a small amount presynaptic inhibition; reciprocal synapses were observed in only 3 of the 18 ribbon synapses. There was no glycinergic crossover inhibition, another local interaction that would enhance light responses. Local postsynaptic inhibition was somewhat more common; in 6 instances, the bipolar cells presynaptic to the gamma cell or their electrically coupled neighbors also provided input to an amacrine cell that inhibited the gamma cell. The other amacrine cell inputs to the gamma cell should have a much greater impact on the light responses because they are far more numerous. These are from axons and long dendrites of GABAergic amacrine cells, and they provide 60% of all the input. This finding suggests that many types of stimuli in the receptive field surround or outside of the classical receptive field would provide potent inhibition to the gamma cell.

Conclusions: The synaptic inputs rsuggest that gamma cells in rabbit retina would have light responses like their homologs in mouse retina, OFF responses to small stimuli in the receptive field center that are suppressed by a variety of larger stimuli. Thus, they would signal object motion selectively.

The Rod-Cone Crossover Connectome of Mammalian Bipolar Cells

We have a new publication out (direct link), The rod-cone crossover connectome of mammalian bipolar cells authored by Scott Lauritzen, Crystal Sigulinsky, James Anderson, Michael Kalloniatis, Noah Nelson, Danny Emrich, Chris Rapp, Nicolas McCarthy, Ethan Kerzner, Mariah Meyer, Bryan W. Jones, and Robert Marc.

Abstract: The basis of cross-suppression between rod and cone channels has long been an enigma. Using rabbit retinal connectome RC1, we show that all cone bipolar cell (BC) classes inhibit rod BCs via amacrine cell (AC) motifs (C1-6); that all cone BC classes are themselves inhibited by AC motifs (R1-5, R25) driven by rod BCs. A sparse symmetric AC motif (CR) is presynaptic and postsynaptic to both rod and cone BCs. ON cone BCs of all classes drive inhibition of rod BCs via motif C1 wide-field GABAergic ACs (γACs) and motif C2 narrow field glycinergic ON ACs (GACs). Each rod BC receives ≈ 10 crossover AC synapses and each ON cone BC can target ≈ 10 or more rod BCs via separate AC processes. OFF cone BCs mediate monosynaptic inhibition of rod BCs via motif C3 driven by OFF γACs and GACs and disynaptic inhibition via motifs C4 and C5 driven by OFF wide-field γACs and narrow-field GACs, respectively. Motifs C4 and C5 form halos of 60-100 inhibitory synapses on proximal dendrites of AI γACs. Rod BCs inhibit surrounding arrays of cone BCs through AII GAC networks that access ON and OFF cone BC patches via motifs R1, R2, R4 R5 and a unique ON AC motif R3 that collects rod BC inputs and targets ON cone BCs. Crossover synapses for motifs C1, C4, C5 and R3 are 3-4x larger than typical feedback synapses, which may be a signature for synaptic winner-take-all switches.

Retinal Remodeling And Metabolic Alterations in Human AMD

We have a new publication out (direct link, open access), Müller Cell Metabolic Chaos During Retinal Degeneration authored by Bryan W. JonesRebecca Pfeiffer, William Ferrell, Carl Watt, James Tucker, and Robert Marc.


Age-related macular degeneration (AMD) is a progressive retinal degeneration resulting in central visual field loss, ultimately causing debilitating blindness. AMD affects 18% of Americans from 65 to 74, 30% older than 74 years of age and is the leading cause of severe vision loss and blindness in Western populations. While many genetic and environmental risk factors are known for AMD, we currently know less about the mechanisms mediating disease progression. The pathways and mechanisms through which genetic and non-genetic risk factors modulate development of AMD pathogenesis remain largely unexplored. Moreover, current treatment for AMD is palliative and limited to wet/exudative forms. Retina is a complex, heterocellular tissue and most retinal cell classes are impacted or altered in AMD. Defining disease and stage-specific cytoarchitectural and metabolic responses in AMD is critical for highlighting targets for intervention. The goal of this article is to illustrate cell types impacted in AMD and demonstrate the implications of those changes, likely beginning in the retinal pigment epithelium (RPE), for remodeling of the the neural retina. Tracking heterocellular responses in disease progression is best achieved with computational molecular phenotyping (CMP), a tool that enables acquisition of a small molecule fingerprint for every cell in the retina. CMP uncovered critical cellular and molecular pathologies (remodeling and reprogramming) in progressive retinal degenerations such as retinitis pigmentosa (RP). We now applied these approaches to normal human and AMD tissues mapping progression of cellular and molecular changes in AMD retinas, including late-stage forms of the disease.

Müller Cell Metabolic Chaos During Retinal Degeneration

We have a new publication out (direct link, open access), Müller Cell Metabolic Chaos During Retinal Degeneration authored by Rebecca PfeifferRobert Marc, Mineo Kondo, Hiroko Terasaki and Bryan W. Jones.


Müller cells play a critical role in retinal metabolism and are among the first cells to demonstrate metabolic changes in retinal stress or disease. The timing, extent, regulation, and impacts of these changes are not yet known. We evaluated metabolic phenotypes of Müller cells in the degenerating retina.

Retinas harvested from wild-type (WT) and rhodopsin Tg P347L rabbits were fixed in mixed aldehydes and resin embedded for computational molecular phenotyping (CMP). CMP facilitates small molecule fingerprinting of every cell in the retina, allowing evaluation of metabolite levels in single cells.

CMP revealed signature variations in metabolite levels across Müller cells from TgP347L retina. In brief, neighboring Müller cells demonstrated variability in taurine, glutamate, glutamine, glutathione, glutamine synthetase (GS), and CRALBP. This variability showed no correlation across metabolites, implying the changes are functionally chaotic rather than simply heterogeneous. The inability of any clustering algorithm to classify Müller cell as a single class in the TgP347L retina is a formal proof of metabolic variability in the present in degenerating retina.

Although retinal degeneration is certainly the trigger, Müller cell metabolic alterations are not a coherent response to the microenvironment. And while GS is believed to be the primary enzyme responsible for the conversion of glutamate to glutamine in the retina, alternative pathways appear to be unmasked in degenerating retina. Somehow, long term remodeling involves loss of Müller cell coordination and identity, which has negative implications for therapeutic interventions that target neurons alone.

Pattern Recognition Reveals Different Visual Field Signature Patterns When Using Spatially Equated Test Sizes Compared To Standard Goldmann III Alone

This abstract was presented today, May 2th at the 2016 Association for Research in Vision and Opthalmology (ARVO) meetings in Seattle, Washington by Michael Kalloniatis, Robert E. Marc, Sieu K. Khuu, Jack Phu, Barbara Zangerl, Lisa Nivison-Smith, Bryan W. Jones, and Rebecca L. Pfeiffer. 

Abstract Number: 4745

Author Block: Michael Kalloniatis, Robert E. Marc, Sieu K. Khuu, Jack Phu, Barbara Zangerl, Lisa Nivison-Smith, Bryan W. Jones, Rebecca L. Pfeiffer
1 Centre for Eye Health, SOVS, University of New South Wales, Kensington, New South Wales, Australia; 2 SOVS, UNSW, Sydney, New South Wales, Australia; 3 Univ of Utah/Moran Eye Center, Salt Lake City, Utah, United States

Disclosure Block:Michael Kalloniatis, 2014/094035 A1 (USA) and 13865419.9 (EU) (Code P (Patent) ); Robert E. Marc, None; Sieu K. Khuu, 2014/094035 A1 (USA) and 13865419.9 (EU) (Code P (Patent) ); Jack Phu, None; Barbara Zangerl, None; Lisa Nivison-Smith, None; Bryan W. Jones, None; Rebecca L. Pfeiffer, None

Purpose:To identify areas within the visual field with matching contrast sensitivity (CS) signature patterns as a function of age using pattern recognition and determine the discrimination of CS data when using spatially equated test stimuli compared to the single size Goldmann (G)III alone.

Methods:52 subjects (classified in decade age groups from 20-60+ years) were tested using the Humphrey Visual Field Analyser 30-2 paradigm in full threshold mode for GI to GV. At least two thresholds were obtained per size. Two visual field maps were analyzed: a spatially equated visual field where GI was used centrally, GII mid-peripherally and GIII in the outer rings to place the test size at or close to complete spatial summation and a second where a single GIII was used at all locations. Thresholds were expressed as dB* (Khuu & Kalloniatis, IOVS 2015), converted to pixel values and analyzed using an unsupervised classification using isodata clustering (PCI, Geomatica, Canada). Class separation was extracted across the ages to develop dot plots of decade measures of CS.

Results:The 77 data points across the central 60° visual field can be distilled into 6 functional classes using the spatially equated visual field (Class separation 1). The 6 classes reflect areas in visual space that change in a similar manner across the ages. The use of the single GIII target resulted in only 4 classes displaying a poorer discrimination over the central visual field (Class separation 2). Extracted dot plots from class separation illustrated average CS within each class could be assessed across the decades.

Conclusions:When using spatially equated visual field testing, concentric areas were separated into distinct CS signatures consistent with known visual field sensitivity. We confirmed these areas change systematically with age. GIII failed to discriminate central areas of the 30-2 that likely reflects the fact that this size operates outside complete spatial summation and thus may not be the optimal test size for assessing visual function in the central visual field. More importantly, we showed pattern recognition can be applied to complex visual field data sets to identify common features and age-related visual function changes. This analysis allows regions to be averaged as they are statistically identical: this approach will likely assist structure-function studies.