MRI-based prostate cancer classification using 3D efficient capsule network.

BACKGROUND: Prostate cancer (PCa) is the most common cancer in men and the second leading cause of male cancer-related death. Gleason score (GS) is the primary driver of PCa risk-stratification and medical decision-making, but can only be assessed at present via biopsy under anesthesia. Magnetic resonance imaging (MRI) is a promising non-invasive method to further characterize PCa, providing additional anatomical and functional information. Meanwhile, the diagnostic power of MRI is limited by qualitative or, at best, semi-quantitative interpretation criteria, leading to inter-reader variability. PURPOSES: Computer-aided diagnosis employing quantitative MRI analysis has yielded promising results in non-invasive prediction of GS. However, convolutional neural networks (CNNs) do not implicitly impose a frame of reference to the objects. Thus, CNNs do not encode the positional information properly, limiting method robustness against simple image variations such as flipping, scaling, or rotation. Capsule network (CapsNet) has been proposed to address this limitation and achieves promising results in this domain. In this study, we develop a 3D Efficient CapsNet to stratify GS-derived PCa risk using T2-weighted (T2W) MRI images. METHODS: In our method, we used 3D CNN modules to extract spatial features and primary capsule layers to encode vector features. We then propose to integrate fully-connected capsule layers (FC Caps) to create a deeper hierarchy for PCa grading prediction. FC Caps comprises a secondary capsule layer which routes active primary capsules and a final capsule layer which outputs PCa risk. To account for data imbalance, we propose a novel dynamic weighted margin loss. We evaluate our method on a public PCa T2W MRI dataset from the Cancer Imaging Archive containing data from 976 patients. RESULTS: Two groups of experiments were performed: (1) we first identified high-risk disease by classifying low + medium risk versus high risk; (2) we then stratified disease in one-versus-one fashion: low versus high risk, medium versus high risk, and low versus medium risk. Five-fold cross validation was performed. Our model achieved an area under receiver operating characteristic curve (AUC) of 0.83 and 0.64 F1-score for low versus high grade, 0.79 AUC and 0.75 F1-score for low + medium versus high grade, 0.75 AUC and 0.69 F1-score for medium versus high grade and 0.59 AUC and 0.57 F1-score for low versus medium grade. Our method outperformed state-of-the-art radiomics-based classification and deep learning methods with the highest metrics for each experiment. Our divide-and-conquer strategy achieved weighted Cohen's Kappa score of 0.41, suggesting moderate agreement with ground truth PCa risks. CONCLUSIONS: In this study, we proposed a novel 3D Efficient CapsNet for PCa risk stratification and demonstrated its feasibility. This developed tool provided a non-invasive approach to assess PCa risk from T2W MR images, which might have potential to personalize the treatment of PCa and reduce the number of unnecessary biopsies.

2D medical image synthesis using transformer-based denoising diffusion probabilistic model.

Objective. Artificial intelligence (AI) methods have gained popularity in medical imaging research. The size and scope of the training image datasets needed for successful AI model deployment does not always have the desired scale. In this paper, we introduce a medical image synthesis framework aimed at addressing the challenge of limited training datasets for AI models.Approach. The proposed 2D image synthesis framework is based on a diffusion model using a Swin-transformer-based network. This model consists of a forward Gaussian noise process and a reverse process using the transformer-based diffusion model for denoising. Training data includes four image datasets: chest x-rays, heart MRI, pelvic CT, and abdomen CT. We evaluated the authenticity, quality, and diversity of the synthetic images using visual Turing assessments conducted by three medical physicists, and four quantitative evaluations: the Inception score (IS), Fréchet Inception Distance score (FID), feature similarity and diversity score (DS, indicating diversity similarity) between the synthetic and true images. To leverage the framework value for training AI models, we conducted COVID-19 classification tasks using real images, synthetic images, and mixtures of both images.Main results. Visual Turing assessments showed an average accuracy of 0.64 (accuracy converging to50%indicates a better realistic visual appearance of the synthetic images), sensitivity of 0.79, and specificity of 0.50. Average quantitative accuracy obtained from all datasets were IS = 2.28, FID = 37.27, FDS = 0.20, and DS = 0.86. For the COVID-19 classification task, the baseline network obtained an accuracy of 0.88 using a pure real dataset, 0.89 using a pure synthetic dataset, and 0.93 using a dataset mixed of real and synthetic data.Significance. A image synthesis framework was demonstrated for medical image synthesis, which can generate high-quality medical images of different imaging modalities with the purpose of supplementing existing training sets for AI model deployment. This method has potential applications in many data-driven medical imaging research.

Incidence and predictors of toxicity in the management of vulvar squamous cell carcinoma treated with radiation therapy.

Purpose/Objective: Given the rarity of vulvar cancer, data on the incidence of acute and late severe toxicity and patients' symptom burden from radiotherapy (RT) are lacking. Materials/Methods: This multi-center, single-institution study included patients with vulvar squamous cell carcinoma treated with curative intent RT between 2009 and 2020. Treatment-related acute and late grade ≥ 3 toxicities and late patient subjective symptoms (PSS) were recorded. Results: Forty-two patients with predominantly stage III/IV disease (n = 25, 59.5 %) were treated with either definitive (n = 25, 59.5 %) or adjuvant (n = 17, 40.5 %) external beam RT to a median dose of 64 Gy and 59.4 Gy, respectively. Five patients received a brachytherapy boost with a median total dose of 84.3 Gy in 2 Gy-equivalent dose (EQD2). Intensity-modulated RT was used in 37 (88.1 %) of patients, and 25 patients (59.5 %) received concurrent chemotherapy. Median follow-up was 27 months. Acute grade ≥ 3 toxicity occurred in 17 patients (40.5 %), including 13 (31.0 %) acute grade 3 skin events. No factors, including total RT dose (p = 0.951), were associated with acute skin toxicity. Eleven (27.5 %) patients developed late grade ≥ 3 toxicity events, including 10 (23.8 %) late grade ≥ 3 skin toxicity events. Patients with late grade ≥ 3 skin toxicity had a higher mean body-mass index (33.0 vs 28.2 kg/m2; p = 0.009). Common late PSS included vaginal pain (n = 15, 35.7 %), skin fibrosis (n = 10, 23.8 %), and requirement of long-term opiates (n = 12, 28.6 %). Conclusion: RT for vulvar cancer is associated with considerable rates of severe acute and late toxicity and PSS burden. Larger studies are needed to identify risk factors, explore toxicity mitigation strategies, and assess patient-reported outcomes.

Contemporary prostate cancer treatment choices in multidisciplinary clinics referenced to national trends.

BACKGROUND: The purpose of this study was to assess treatment choices among men with prostate cancer who presented at The University of Texas MD Anderson Cancer Center multidisciplinary (MultiD) clinic compared with nationwide trends. METHODS: In total, 4451 men with prostate cancer who presented at the MultiD clinic from 2004 to 2016 were analyzed. To assess nationwide trends, the authors analyzed 392,710 men with prostate cancer who were diagnosed between 2004 and 2015 from the Surveillance, Epidemiology, and End Results (SEER) database. The primary endpoint was treatment choice as a function of pretreatment demographics. RESULTS: Univariate analyses revealed similar treatment trends in the MultiD and SEER cohorts. The use of procedural forms of definitive therapy decreased with age, including brachytherapy and prostatectomy (all P < .05). Later year of diagnosis/clinic visit was associated with decreased use of definitive treatments, whereas higher risk grouping was associated with increased use (all P < .001). Patients with low-risk disease treated at the MultiD clinic were more likely to receive nondefinitive therapy than patients in SEER, whereas the opposite trend was observed for patients with high-risk disease, with a substantial portion of high-risk patients in SEER not receiving definitive therapy. In the MultiD clinic, African American men with intermediate-risk and high-risk disease were more likely to receive definitive therapy than white men, but for SEER the opposite was true. CONCLUSIONS: Presentation at a MultiD clinic facilitates the appropriate disposition of patients with low-risk disease to nondefinitive strategies of patients with high-risk disease to definitive treatment, and it may obviate the influence of race.

The PEX1 ATPase Stabilizes PEX6 and Plays Essential Roles in Peroxisome Biology.

A variety of metabolic pathways are sequestered in peroxisomes, conserved organelles that are essential for human and plant survival. Peroxin (PEX) proteins generate and maintain peroxisomes. The PEX1 ATPase facilitates recycling of the peroxisome matrix protein receptor PEX5 and is the most commonly affected peroxin in human peroxisome biogenesis disorders. Here, we describe the isolation and characterization of, to our knowledge, the first Arabidopsis (Arabidopsis thaliana) pex1 missense alleles: pex1-2 and pex1-3pex1-2 displayed peroxisome-related defects accompanied by reduced PEX1 and PEX6 levels. These pex1-2 defects were exacerbated by growth at high temperature and ameliorated by growth at low temperature or by PEX6 overexpression, suggesting that PEX1 enhances PEX6 stability and vice versa. pex1-3 conferred embryo lethality when homozygous, confirming that PEX1, like several other Arabidopsis peroxins, is essential for embryogenesis. pex1-3 displayed symptoms of peroxisome dysfunction when heterozygous; this semidominance is consistent with PEX1 forming a heterooligomer with PEX6 that is poisoned by pex1-3 subunits. Blocking autophagy partially rescued PEX1/pex1-3 defects, including the restoration of normal peroxisome size, suggesting that increasing peroxisome abundance can compensate for the deficiencies caused by pex1-3 and that the enlarged peroxisomes visible in PEX1/pex1-3 may represent autophagy intermediates. Overexpressing PEX1 in wild-type plants impaired growth, suggesting that excessive PEX1 can be detrimental. Our genetic, molecular, and physiological data support the heterohexamer model of PEX1-PEX6 function in plants.

The Roles of ß-Oxidation and Cofactor Homeostasis in Peroxisome Distribution and Function in Arabidopsis thaliana.

Key steps of essential metabolic pathways are housed in plant peroxisomes. We conducted a microscopy-based screen for anomalous distribution of peroxisomally targeted fluorescence in Arabidopsis thaliana This screen uncovered 34 novel alleles in 15 genes affecting oil body mobilization, fatty acid ß-oxidation, the glyoxylate cycle, peroxisome fission, and pexophagy. Partial loss-of-function of lipid-mobilization enzymes conferred peroxisomes clustered around retained oil bodies without other notable defects, suggesting that this microscopy-based approach was sensitive to minor perturbations, and that fatty acid ß-oxidation rates in wild type are higher than required for normal growth. We recovered three mutants defective in PECTIN METHYLESTERASE31, revealing an unanticipated role in lipid mobilization for this cytosolic enzyme. Whereas mutations reducing fatty acid import had peroxisomes of wild-type size, mutations impairing fatty acid ß-oxidation displayed enlarged peroxisomes, possibly caused by excess fatty acid ß-oxidation intermediates in the peroxisome. Several fatty acid ß-oxidation mutants also displayed defects in peroxisomal matrix protein import. Impairing fatty acid import reduced the large size of peroxisomes in a mutant defective in the PEROXISOMAL NAD+ TRANSPORTER (PXN), supporting the hypothesis that fatty acid accumulation causes pxn peroxisome enlargement. The diverse mutants isolated in this screen will aid future investigations of the roles of ß-oxidation and peroxisomal cofactor homeostasis in plant development.

Changes in interhelical hydrogen bonding upon rhodopsin activation.

Hydrogen bonding interactions between transmembrane helices stabilize the visual pigment rhodopsin in an inactive conformation in the dark. The crystal structure of rhodopsin has previously revealed that Glu122 and Trp126 on transmembrane helix H3 form a complex hydrogen bonding network with Tyr206 and His211 on H5, while the indole nitrogen of Trp265 on H6 forms a water-mediated hydrogen bond with Asn302 on H7. Here, we use solid-state magic angle spinning NMR spectroscopy to probe the changes in hydrogen bonding upon rhodopsin activation. The NMR chemical shifts of 15N-labeled tryptophan are consistent with the indole nitrogens of Trp126 and Trp265 becoming more weakly hydrogen bonded between rhodopsin and metarhodopsin II. The NMR chemical shifts of 15N-labeled histidine show that His211 is neutral; the unprotonated imidazole nitrogen is not coordinated to zinc in rhodopsin and becomes more strongly hydrogen bonded in metarhodopsin II. Moreover, measurements of rhodopsin containing 13C-labeled histidine show that a strong hydrogen bond between the side-chain of Glu122 and the backbone carbonyl of His211 is disrupted in metarhodopsin II. The implications of these observations for the activation mechanism of rhodopsin are discussed.

Predictors of acute esophagitis in patients with non-small-cell lung carcinoma treated with concurrent chemotherapy and hyperfractionated radiotherapy followed by surgery.

PURPOSE: To evaluate possible clinical and dosimetric predictors of acute esophagitis in patients with locally advanced non-small-cell lung carcinoma treated in a prospective Phase I-II trimodality protocol. METHODS AND MATERIALS: The data from 36 patients with Stage III non-small-cell lung carcinoma treated in a Phase I-II high-dose concurrent chemoradiotherapy protocol were analyzed for possible predictors of acute esophagitis. The median age was 58 years (range, 38-77 years). Patients included in this study had either Stage IIIA (n = 24) or IIIB (n = 12) disease. All patients were treated with induction concurrent carboplatin (area under the plasma concentration-time curve 1), vinorelbine (5-15 mg/m(2)), and hyperfractionated radiotherapy (69.6 Gy) followed by consolidation chemotherapy (carboplatin area under the plasma concentration-time curve 6, vinorelbine 25 mg/m(2), docetaxel 75 mg/m(2)) or surgery (n = 19) plus consolidation chemotherapy. Acute toxicities were graded using the Radiation Therapy Oncology Group criteria. The following clinical and dosimetric parameters were analyzed: age, gender, race, T stage, N stage, pretreatment body mass index, percentage of weight lost during therapy, pretherapy serum albumin, tumor location, length of esophagus in treatment field, percentage of esophagus volume treated to >40, >45, >50, >55, >60, and >65 Gy. These parameters were coded and analyzed against Grade 2 and worse esophagitis using univariate and multivariate regression analyses. RESULTS: Of the 36 patients, Grade 1, 2, and 3 acute esophagitis was observed in 16 (44%), 12 (33%), and 2 (5.5%) patients, respectively. Grade 4 or 5 toxicity was not observed in this patient cohort. Only the pretreatment body mass index (rho = -0.431, p = 0.004) and percentage of esophagus volume treated to >50 Gy (rho = 0.297, p = 0.040) demonstrated a statistically significant correlation with the incidence of Grade 2 or worse esophagitis on univariate analysis. These parameters retained their statistical significance on multivariate regression analysis (p = 0.029 and 0.049, respectively). CONCLUSION: In patients undergoing concurrent high-dose chemotherapy and hyperfractionated radiotherapy, a low pretherapy body mass index and percentage of esophagus volume treated to >50 Gy were significantly associated with acute Grade 2 or worse esophagitis.

Pulmonary resection after curative intent radiotherapy (>59 Gy) and concurrent chemotherapy in non-small-cell lung cancer.

BACKGROUND: Pulmonary resection after chemotherapy and concurrent full-dose radiotherapy (>59 Gy) has previously been associated with unacceptably high morbidity and mortality. Subsequently neoadjuvant therapy protocols have used reduced and potentially suboptimal radiotherapy doses of 45 Gy. We report a series of 40 patients with locally advanced non-small-cell lung cancer who successfully underwent pulmonary resection after receiving greater than 59 Gy radiation and concurrent chemotherapy. Operative results and midterm survival follow-up are presented. METHODS: Data were reviewed from 40 consecutive patients who underwent lung resection after receiving high-dose radiotherapy and concurrent platinum-based chemotherapy between January 1994 and May 2000. The follow-up closing interval for this study was until August 2003 or time of death. RESULTS: Preoperative stage was IIb (7 patients), IIIA (21 patients), IIIB (10 patients), and IV (2 patients with isolated brain metastasis). Thirteen patients exhibited Pancoast tumors. Median time from completion of induction therapy to surgery was 53 days. Twenty-nine lobectomies and 11 pneumonectomies (7 right, 4 left) were performed. There were no postoperative deaths. Intercostal muscle flaps were used prophylactically in all but one pneumonectomy patient. Seven patients required perioperative transfusions. Median intensive care unit (ICU) time averaged 2 days and the total length of stay was 6 days. One patient exhibited postpneumonectomy pulmonary edema and a bronchopleural fistula developed in another patient (not receiving an intercostal muscle flap). Thirty-four of 40 patients (85%; 95% CI: 70%-94%) were downstaged pathologically, 33 out of 40 patients (82.5%, 95% confidence interval [CI]: 67%-93%) indicated no residual lymphadenopathy, and 18 out of 40 patients (45%, 95% CI: 29%-61%) exhibited a complete pathologic response. Median follow-up was 2.8 years. The 1-, 2-, and 5-year overall survival rates were 92.4%, 66.7%, and 46.2%, respectively. Disease-free 1-, 2-, and 5-year survival rates were 73.0%, 67.2%, and 56.4%, respectively. Median disease-free survival has not been reached. CONCLUSIONS: Pulmonary resection may be performed safely after curative intent concurrent chemotherapy and radiotherapy to greater than 59 Gy. High pathologic complete response rates and sterilization of mediastinal lymph nodes were observed accompanied by highly favorable survival rates. This experience, though promising, will require confirmation in a prospective multiinstitutional clinical trial.

Coupling of retinal isomerization to the activation of rhodopsin.

Activation of the visual pigment rhodopsin is caused by 11-cis to -trans isomerization of its retinal chromophore. High-resolution solid-state NMR measurements on both rhodopsin and the metarhodopsin II intermediate show how retinal isomerization disrupts helix interactions that lock the receptor off in the dark. We made 2D dipolar-assisted rotational resonance NMR measurements between (13)C-labels on the retinal chromophore and specific (13)C-labels on tyrosine, glycine, serine, and threonine in the retinal binding site of rhodopsin. The essential aspects of the isomerization trajectory are a large rotation of the C20 methyl group toward extracellular loop 2 and a 4- to 5-A translation of the retinal chromophore toward transmembrane helix 5. The retinal-protein contacts observed in the active metarhodopsin II intermediate suggest a general activation mechanism for class A G protein-coupled receptors involving coupled motion of transmembrane helices 5, 6, and 7.

Dipolar assisted rotational resonance NMR of tryptophan and tyrosine in rhodopsin.

Two dimensional (2D) solid-state (13)C.(13)C dipolar recoupling experiments are performed on a series of model compounds and on the visual pigment rhodopsin to establish the most effective method for long range distance measurements in reconstituted membrane proteins. The effects of uniform labeling, inhomogeneous B(1) fields, relaxation and dipolar truncation on cross peak intensity are investigated through NMR measurements of simple amino acid and peptide model compounds. We first show that dipolar assisted rotational resonance (DARR) is more effective than RFDR in recoupling long-range dipolar interactions in these model systems. We then use DARR to establish (13)C-(13)C correlations in rhodopsin. In rhodopsin containing 4'-(13)C-Tyr and 8,19-(13)C retinal, we observe two distinct tyrosine-to-retinal correlations in the DARR spectrum. The most intense cross peak arises from a correlation between Tyr268 and the retinal 19-(13)CH(3), which are 4.8 A apart in the rhodopsin crystal structure. A second cross peak arises from a correlation between Tyr191 and the retinal 19-(13)CH(3), which are 5.5 A apart in the crystal structure. These data demonstrate that long range (13)C em leader (13)C correlations can be obtained in non-crystalline integral membrane proteins reconstituted into lipid membranes containing less than 150 nmoles of protein. In rhodopsin containing 2-(13)C Gly121 and U-(13)C Trp265, we do not observe a Trp-Gly cross peak in the DARR spectrum despite their close proximity (3.6 A) in the crystal structure. Based on model compounds, the absence of a (13)C em leader (13)C cross peak is due to loss of intensity in the diagonal Trp resonances rather than to dipolar truncation.

Helix packing moments reveal diversity and conservation in membrane protein structure.

Helical membrane proteins are more tightly packed and the packing interactions are more diverse than those found in helical soluble proteins. Based on a linear correlation between amino acid packing values and interhelical propensity, we propose the concept of a helix packing moment to predict the orientation of helices in helical membrane proteins and membrane protein complexes. We show that the helix packing moment correlates with the helix interfaces of helix dimers of single pass membrane proteins of known structure. Helix packing moments are also shown to help identify the packing interfaces in membrane proteins with multiple transmembrane helices, where a single helix can have multiple contact surfaces. Analyses are described on class A G protein-coupled receptors (GPCRs) with seven transmembrane helices. We show that the helix packing moments are conserved across the class A family of GPCRs and correspond to key structural contacts in rhodopsin. These contacts are distinct from the highly conserved signature motifs of GPCRs and have not previously been recognized. The specific amino acid types involved in these contacts, however, are not necessarily conserved between subfamilies of GPCRs, indicating that the same protein architecture can be supported by a diverse set of interactions. In GPCRs, as well as membrane channels and transporters, amino acid residues with small side-chains (Gly, Ala, Ser, Cys) allow tight helix packing by mediating strong van der Waals interactions between helices. Closely packed helices, in turn, facilitate interhelical hydrogen bonding of both weakly polar (Ser, Thr, Cys) and strongly polar (Asn, Gln, Glu, Asp, His, Arg, Lys) amino acid residues. We propose the use of the helix packing moment as a complementary tool to the helical hydrophobic moment in the analysis of transmembrane sequences.

NMR spectroscopy of phosphorylated wild-type rhodopsin: mobility of the phosphorylated C-terminus of rhodopsin in the dark and upon light activation.

Binding of arrestin to light-activated rhodopsin involves recognition of the phosphorylated C-terminus and several residues on the cytoplasmic surface of the receptor. These sites are in close proximity in dark, unphosphorylated rhodopsin. To address the position and mobility of the phosphorylated C-terminus in the active and inactive receptor, we combined high-resolution solution and solid state NMR spectroscopy of the intact mammalian photoreceptor rhodopsin in detergent micelles as a function of temperature. The (31)P NMR resonance of rhodopsin phosphorylated by rhodopsin kinase at the C-terminal tail was observable with single pulse excitation using magic angle spinning until the sample temperature reached -40 degrees C. Below this temperature, the (31)P resonance broadened and was only observable using cross polarization. These results indicate that the phosphorylated C-terminus is highly mobile above -40 degrees C and immobilized at lower temperature. To probe the relative position of the immobilized phosphorylated C-terminus with respect to the cytoplasmic domain of rhodopsin, (19)F labels were introduced at positions 140 and 316 by the reaction of rhodopsin with 2,2,2-trifluoroethanethiol (TET). Solid state rotational-echo double-resonance (REDOR) NMR was used to probe the internuclear distance between the (19)F and the (31)P-labels. The REDOR technique allows (19)F...(31)P distances to be measured out to approximately 12 A with high resolution, but no significant dephasing was observed in the REDOR experiment in the dark or upon light activation. This result indicates that the distances between the phosphorylated sites on the C-terminus and the (19)F sites on helix 8 (Cys 316) and in the second cytoplasmic loop (Cys140) are greater than 12 A in phosphorylated rhodopsin.

A detailed examination of the difference between planned and treated margins in 125I permanent prostate brachytherapy.

PURPOSE: To determine whether potential extraprostatic extension (EPE) of prostate adenocarcinoma is covered by the prescribed dose when permanent 125I implants are planned with a 5-mm treatment margin. METHODS AND MATERIALS: The postimplant dosimetry of 60 consecutive 125I prostate implants was analyzed to determine whether there was a residual 3-mm margin, adequate for treatment of potential EPE. The implants were peripherally loaded and planned with a nominal 5-mm symmetric dose margin. Extraprostatic seeds were not used at midgland, although extraprostatic seeds were placed at the base and apex. The radial distance between the edge of the prostate and the prescription isodose line (145 Gy) was measured at the left lateral, left posterolateral, posterior, right posterolateral, and right lateral positions at the base, midgland, and apex in both the preplan and postimplant dosimetry. RESULTS: The mean postimplant margin at the base (4 +/- 2 mm) was significantly less (p < 0.01) than planned (6 +/- 2 mm) by 2 mm. The planned and postimplant margins at the midgland (5 +/- 1 mm and 5 +/- 2 mm) and apex (7 +/- 2 mm and 7 +/- 3 mm) were indistinguishable (p = 0.31 and p = 0.69, respectively). At the base, 69% of the measurements were > or = 3 mm compared with 89% and 91% at the midgland and apex, respectively. Overall, 83% of the margin measurements were > or = 3 mm. The prostate postimplant V100 and D90 were 96 +/- 4% and 193 +/- 27 Gy, respectively. CONCLUSIONS: A 5-mm planning dose margin can potentially treat most EPE. However, the postimplant margin, like other dosimetric parameters, is sensitive to source placement errors and the percentage of EPE treated depends upon how well the plan is executed. Because implant quality is operator dependent, we would not recommend brachytherapy alone for patients with a high risk of EPE.

Comparison of helix interactions in membrane and soluble alpha-bundle proteins.

Helix-helix interactions are important for the folding, stability, and function of membrane proteins. Here, two independent and complementary methods are used to investigate the nature and distribution of amino acids that mediate helix-helix interactions in membrane and soluble alpha-bundle proteins. The first method characterizes the packing density of individual amino acids in helical proteins based on the van der Waals surface area occluded by surrounding atoms. We have recently used this method to show that transmembrane helices pack more tightly, on average, than helices in soluble proteins. These studies are extended here to characterize the packing of interfacial and noninterfacial amino acids and the packing of amino acids in the interfaces of helices that have either right- or left-handed crossing angles, and either parallel or antiparallel orientations. We show that the most abundant tightly packed interfacial residues in membrane proteins are Gly, Ala, and Ser, and that helices with left-handed crossing angles are more tightly packed on average than helices with right-handed crossing angles. The second method used to characterize helix-helix interactions involves the use of helix contact plots. We find that helices in membrane proteins exhibit a broader distribution of interhelical contacts than helices in soluble proteins. Both helical membrane and soluble proteins make use of a general motif for helix interactions that relies mainly on four residues (Leu, Ala, Ile, Val) to mediate helix interactions in a fashion characteristic of left-handed helical coiled coils. However, a second motif for mediating helix interactions is revealed by the high occurrence and high average packing values of small and polar residues (Ala, Gly, Ser, Thr) in the helix interfaces of membrane proteins. Finally, we show that there is a strong linear correlation between the occurrence of residues in helix-helix interfaces and their packing values, and discuss these results with respect to membrane protein structure prediction and membrane protein stability.