PTx is also described during Valsalva manoeuvres in very regular

PTx is also described during Valsalva manoeuvres in very regular long term marijuana smokers.13 and 14 Our patient had apical bullous lung disease and the fact that his drain was still bubbling after several days on suction suggested he had an ongoing air leak. He declined to have a larger bore drain inserted and unfortunately his drain dislodged prior to full radiological selleck screening library resolution. A referral to cardiothoracic services was considered but as the patient was clinically and radiologically stable following the removal of his drain he was discharged after strong counselling not to fly, dive and to give up smoking completely. However, he

has failed to keep his appointments for follow up. Marijuana is the most common illegal drug used in the UK. There are no conflicts of interest in this paper. “
“A 53 year-old woman with a 3.7-year history of progressive lower-limb weakness

due to amyotrophic lateral sclerosis (ALS), confirmed two years previously, was admitted with a one-month history of episodes of dizziness, some of which were associated with brief loss of consciousness. She was noted to be pale during these events and at least one episode occurred on laughing. She was incontinent during a few Cell Cycle inhibitor events, but always fully recovered within minutes. She had a cough productive of green sputum for one week and on examination she was drowsy with poor respiratory effort. She was wheelchair bound with global flaccid most weakness in the lower limbs, mild upper limb weakness and very mild bulbar impairment. Arterial blood gases showed acute on chronic type-two respiratory failure (pH = 7.17, PaCO2 = 15.1 kPa) and her chest radiograph showed bibasal atelectasis. Non-invasive ventilation (NIV) was initiated together with a seven-day course of amoxicillin. She improved clinically and physiologically with rapid correction of the respiratory acidosis. However, she

suffered profound bradycardia, sometimes associated with transient loss of consciousness, on each occasion the NIV mask was removed (Fig. 1), in the early stages after initiation of NIV. The episodes of bradycardia resolved when NIV was recommenced. She was taking a number of medications that could potentially induce bradycardia: atenolol, diltiazem, ranitidine (cimetidine has been shown to cause bradycardia), and quinine. These were discontinued, but the frequency and severity of the episodes of bradycardia were unaffected. The episodes of bradycardia occurred too rapidly for hypoxia to be implicated as the cause and they persisted after correction of the initial respiratory acidosis. There was no evidence of myocardial infarction or an intrinsic conduction abnormality on her ECG. The episodes of bradycardia were fully blocked by pre-treatment with atropine before removal of the mask (Fig. 1). Subsequently, an isoprenaline infusion was commenced with similar efficacy.

To investigate the reactivity of these systems, gallic acid (GA)

To investigate the reactivity of these systems, gallic acid (GA) was used as a model system for the polyphenols present in foods products (Chvátalová et al., 2008 and Fazary et al., 2009). The formation of the Fe3+–GA complex can be followed over time using spectrophotometry, as the complex has a dark small molecule library screening blue colour (Chvátalová et al., 2008 and Mellican et al., 2003). This increase in absorption was used as an indication for the reactivity of the iron contained in the particles. However, the analysis is complicated by the ability of polyphenols to reduce Fe3+, resulting

in a Fe2+–quinone complex that is also blue. Although various possible pathways are known for this reaction (Arif Kazmi et al., 1987, Funabiki et al., 1986 and Powell and Taylor, 1982), the most probable one under physiological conditions is described by Hynes (2001). Once the quinone has been formed, the Fe2+ can be oxidised to form a new complex with free gallic acid. As will be shown here, the oxidation reaction is much slower than the initial complex formation and the cyclisation of the reaction MAPK inhibitor can be limited

by sealing the sample air tight. The difference between the two complexes can be distinguished using spectrophotometry, since they have different absorption maxima, although it does interfere with the quantification of the complexation reaction. Due to the side reactions and the complexity of the system, only the initial reactivity during the first 5 h after addition was analysed and only qualitative comparisons between identically prepared samples were made. FeCl3·6H2O (ACS reagent grade, 97%) and zein protein were obtained from Sigma Aldrich. Na4P2O7·10H2O (ACS reagent grade), CaCl2·2H2O (ACS reagent grade, ⩾99%) and NaCl (p.a., ⩾99.5%) were purchased from Merck and MgCl2·6H2O (puriss. p.a., ⩾99%) from Fluka. Gallic acid (extra pure, ⩾99.5%) was obtained from Scharlau Chemie. All

chemicals were used as received; aqueous solutions were prepared using water deionised by a Millipore Synergy water purification system. Systems were dialysed using Spectra/Por 2 Dialysis Membrane, molecular weight cut-off (MWCO) 12–14 Da, corresponding to roughly a 1.5 nm pore size. Iron pyrophosphate was prepared as described previously (van Leeuwen et al., 2012a and van Leeuwen Phosphoribosylglycinamide formyltransferase et al., 2012c). Briefly, nanoparticles were prepared by coprecipitation of Na4P2O7 with FeCl3. 0.86 mmol iron chloride dissolved in 50 ml water was added drop wise, over about 15 min to 0.64 mmol sodium pyrophosphate in 100 ml. A turbid white precipitate formed in the final 5 min of the addition (van Leeuwen et al., 2012c), the resulting dispersion had a pH of 4. The pH-dependent preparation comprised of two steps: first, the precipitation and washing of the intermediate pyrophosphate salt, which was subsequently redissolved in acid and then precipitated in an alkaline solution. For the intermediate, 50 ml 1 M M2+ Cl2 solution was added drop wise over 1 h to 800 ml 0.

In Cheddar cheese, the peptide αs1-CN f1–23 is further


In Cheddar cheese, the peptide αs1-CN f1–23 is further

hydrolysed by proteinases from Lactococcus lactis ssp. cremoris into smaller peptides, which present bitter taste ( Singh et al., 2003). Since Prato cheese is also made with this starter culture, it is probable that this hydrolysis also occurs, affecting TCA 12%-SN. Thus, pH 4.6-SN and TCA 12%-SN in Prato cheese were essentially affected by residual chymosin, plasmin and by proteolytic enzymes of lactic acid bacteria. According to Sousa et al. (2001), since proteolysis is one of the Selleckchem R428 main biochemical events during the ripening of cheese, it is desirable to include a general assay for proteolysis, such as the determination of soluble N as % of total N, as has been done. If the objectives of the study cover investigating the effect of one of the agents of proteolysis in cheese, such as different types of coagulants, the methodology should be chosen so as to emphasise the level of proteolysis caused by that agent. In this case, for example, proteolysis evolution could be followed by urea–polyacrylamide gel electrophoresis (urea–PAGE) and the peptide profiles of the pH 4.6-soluble fraction should be determined by reverse phase-high performance liquid chromatography (RP-HPLC). Therefore, click here proteolysis was assayed by the frequently used method of monitoring casein proteolytic processes:

polyacrylamide Ribonucleotide reductase gel electrophoresis using urea, which makes possible to visualise the integrity of casein fractions during ripening (Fig. 2). In Fig. 2A, two main casein groups were identified in the urea–PAGE: αs1-casein, with higher electrophoretic mobility and β-casein, with lower mobility (Silva & Malcata, 2004). The region of family αs2-casein can also be seen, whose electrophoretic

mobilities is between caseins αs1 and β (Sgarbieri, 2005). Fig. 2B shows casein degradation in cheeses made with commercial coagulant (H1–H60) and with coagulant from T. indicae-seudaticae N31 (T1–T60) during 60 days of ripening. Degradation of αs1-casein is seen, more pronounced in cheeses made with commercial coagulant, showing that the hydrolysis of casein molecules is specific for the type of coagulant used ( Lawrence et al., 1987). Degradation of β-casein is also seen, more intense in cheeses made with coagulant from T. indicae-seudaticae N31, with formation of its hydrolysis products, the γ-caseins, which accumulate in cheese ( Singh et al., 2003). Plasmin is one of the agents responsible for the proteolysis during cheese ripening acting especially in the initial stages along with residual coagulant, liberating peptides which will serve as substrate for proteinases from starter and non starter bacteria ( Fox, 1989 and Visser, 1993). Besides plasmin, chymosin also acts on β-casein, on the bond between Leu192 and Tyr193 ( Visser, 1993). It can also be seen in Fig.

IL-1β levels of the liver tissue in the probiotics and KRG groups

IL-1β levels of the liver tissue in the probiotics and KRG groups decreased compared with check details those in the alcohol group. These results

match those of earlier studies, in which Rg3, an ingredient of Panax ginseng active in neural stem cells, attenuated the upregulation of the LPS-induced IL-1β level [24]. In addition, ginsenoside Rd pretreatment attenuated the increased expression of proinflammatory cytokines (e.g., IL-1β and TNF-α) due to lead (Pb) exposure [25]. Another study demonstrated the efficacy of probiotics in lowering the heightened IL-1β level induced by Candida albicans infection in mice [26]. Assuming IL-1β to be a dangerous cytokine in the ALD inflammatory cascade, Lacidofil and

KRG may be hypothesized to have potent anti-ALD effects. We used a chronic ethanol feeding model (the NIAAA model for ALD). The 4–6-week Lieber–DeCarli diet containing ethanol has been widely used by many laboratories. However, this model induces only mild steatosis and elevates serum alanine aminotransferase slightly, with little or no liver inflammation [27]. In our study, in the Sorafenib price liver function test, there was no significant change as shown in the NIAAA model for ALD. The pathological findings of our study showed that alcohol-induced steatosis was significantly reduced by KRG and urushiol. In addition, two mice developed Grade 2 steatosis in the alcohol group and one in the KRG group. Therefore, it is supposed that KRG and urushiol can be used in the treatment of ALD, especially steatosis of liver. Of the agents that we evaluated, KRG was notably the

most effective in reducing the molecular markers that we assessed in mice. However, because the sample size was limited, serum levels of TLR-4, IL-1β, and TNF-α were not significantly ameliorated. Furthermore, injurious cytokines such as IL-6 were not assessed in this study. Therefore, additional clinical or animal model studies are needed. In conclusion, the current study suggests that KRG, urushiol, and probiotics have potential therapeutic effects, which (in the context of ALD) implicates immune-modulated cytokines in the TLR-4 pathway. All contributing authors declare no conflicts of interest. This research was supported Glycogen branching enzyme by a grant from the Korea Society of Ginseng (2011); the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Education, Science, and Technology (NRF-2010-0021482); Cooperative Research Program for Agriculture Science and Technology Development (Project No. PJ009859), Rural Development Administration, Republic of Korea; and Hallym University Research Fund. “
“The Picornaviridae are currently divided into nine genera, three of which (Hepatoviruses, Rhinoviruses, and Enteroviruses) are causative agents of human diseases [1].

In the spatial Stroop task, the presence of response


In the spatial Stroop task, the presence of response

conflict had less of an additional effect on the dominant task over and above the effect of interruptions. One possible reason for this divergence is that in order to avoid contingencies between the irrelevant and the relevant dimensions (e.g., Melara & Algom, 2003) locations and word stimuli were selected randomly, leading to an average conflict probability of p = .75. In situations with high probability of conflict, conflict effects are often reduced, possibly Selleck BMS754807 because of a general tightening of control ( Tzelgov et al., 1992). Another reason for a relatively small contribution of conflict trials to the cost asymmetry is that in this experiment, conflict effects were “diluted” across RTs and errors, whereas in the endogenous/exogenous task conflict effects affected RTs only. The error learn more effects we did observe in Experiment 5 were in the same direction as RT effects, albeit only approaching the statistical significance criterion. Our goal was to explore the conditions that make it difficult to select between competing control settings, specifically between endogenous and exogenously controlled

attention. On a theoretical level we started out by proposing two conditions that have to be met so that subjects experience substantial selection costs. First, LTM needs to contain memory traces from earlier selection instances with the competing task. Second, these memory traces produce interference only once working memory is forced from a maintenance to an updating state, such as through strong bottom-up interference (as in the endogenous task), task switches, or during

recovery from externally imposed interruptions (as in the exogenous task during post-interruption trials). From these assumptions we derived and empirically confirmed the prediction that the asymmetric selection costs (i.e., larger costs for dominant Adenosine than for non-dominant control settings) arise after any interruption of ongoing processing. This is in contrast to predictions from the currently dominant “carry-over” account of task switching (e.g., Gilbert & Shallice, 2002), which cannot explain a selection cost asymmetry that arises from mere interruptions. Rather, for this account the trial-to-trial clash between competing task or control settings is a necessary condition for the selection cost asymmetry to arise. Cost asymmetries in the absence of task switches between competing tasks have been reported occasionally in the past (Allport and Wylie, 2000 and Bryck and Mayr, 2008). The current results go significantly beyond the existing evidence and allow us to both strengthen and fill in important details about our rather broad starting propositions: First, the interruption-based cost asymmetry is fairly general. In particular, it occurs both in situations in which subjects need to select between competing attentional control settings (Exp. 1–4) and between competing stimulus–response mappings (Exp.

(i) The effect of environmental characteristics (distance to seed

(i) The effect of environmental characteristics (distance to seed source, % vascular plant cover, % woody debris, altitude and soil pH) click here on the tree regeneration densities

were examined using Spearman rank correlation coefficients. The analyses were carried out separately for the dominant species that were identified (birch, alder, rowan, willow and oak). Ground flora characteristics in each quadrat were analysed as: (i) Total number of species, S, (ii) % vascular plant cover of each species, and (iii) linear regression analysis was used to examine the difference in vascular plant coverage with time since clearfelling. A total of 14 tree and shrub species were found to be regenerating, of which 10 were species native to Great Britain. The non-native species consisted of three conifers (Sitka spruce, Pinus contorta (lodgepole pine) and larch) and one broadleaved species (Alnus

incana (grey alder)). The native species were birch, oak, buy BMS-387032 rowan, willow, common alder, Fraxinus excelsior (ash), holly, Fagus sylvatica (common beech), Corylus avellana (common hazel) and Juniperus communis (common juniper). The mean density of regeneration of native species on clearfelled sites varied from 0 stems/ha to >5000 stems/ha ( Table 2). While the regeneration density of non-native tree species is shown in Table 2 it is important to note that in a number of study sites regenerating non-native conifers had been felled, making it difficult to draw any conclusions about the frequency of non-native regeneration. The linear regression of time since clearfelling on regeneration density of native species was not found

to be significant (r2 = 0.26, n.s.). Table 3 shows the density of regeneration for native species and the fraction of clearfelled sites where each species was recorded. Regeneration was dominated by birch and rowan. Whilst the regeneration of holly and oak were recorded infrequently (<20% of sites), relatively high regeneration densities were recorded at specific sites for these species (for example, 723 stems/ha in the case of oak). The regeneration density of birch and alder was found to be negatively correlated with SB-3CT distance from seed source (see Table 4). In the case of birch, for example, 63% of regeneration occurred within 20 m of a seed source. No significant relationship was found for rowan or oak. No significant relationship between plant cover and regeneration density was seen for any species. However, when the regenerating trees were divided into sapling (taller than 0.5 m) or seedling (shorter than 0.5 m) categories then a significant negative correlation was seen between birch seedling density and vascular plant cover. Birch also showed a significant negative correlation with the percentage of brash (woody debris). No such effects were noted for alder, willow, oak or rowan.

The same results were

The same results were click here obtained when

the treatments were performed at 6 h p.i. (data not shown). Nevertheless, Su et al. (2008) demonstrated that UL52 (β) and UL13 (γ) mRNA levels were inhibited by digitoxin. The β-actin (Fig. 5D), was used as an internal standard, and the expression level of its mRNA was not affected. Since the glucoevatromonoside did not inhibit mRNA expression, the next step of HSV replication to be evaluated was the protein synthesis. Likewise glucoevatromonoside, acyclovir, furosemide + potassium chloride (KCl) as well as their combinations were also tested. The relevance of intracellular K+ to the viral replication has already been reported (Hartley et al., 2006, Hartley et al., 1993 and Nagai et al., 1972). Furosemide is a loop diuretic also known as an inhibitor of Na+K+Cl− cotransport activity (NKCC), which prevents the entry of K+ into the cells (Russel, 2000), and has also shown antiherpes activity (Hartley et al., 2006). Thus, furosemide was investigated in order to check if it was able to reduce the viral protein levels. Likewise, the supplementation of LGK-974 solubility dmso K+ by adding KCl to the culture medium was also tested to confirm if this ion was important for the viral inhibition caused by glucoevatromonoside. Fig. 6 shows

the effects of these treatments on some HSV-1 proteins and on β-actin that is use as an internal standard. As shown in Fig. 6, β-actin cell protein was expressed in all treatments; consequently, the tested drugs were not cytotoxic. The individual treatments with glucoevatromonoside (lane 5) and acyclovir (lane 3) reduced the levels of all tested viral proteins, when compared to virus control. The glucoevatromonoside completely inhibited all viral protein synthesis, whereas acyclovir was able to inhibit completely only the UL42 and the gB proteins expression. The treatment with furosemide (lane 8) did not reduce the levels of any viral protein, when compared to viral control indicating that this drug Acetophenone could not affect

this stage of HSV-1 replication or that the tested concentration was insufficient to induce protein synthesis inhibition. When the treatment was performed with glucoevatromonoside + acyclovir (lane 4) or glucoevatromonoside + furosemide (lane 9), a complete inhibition of protein levels was also detected, as well as when glucoevatromonoside (lane 5) was tested alone. Therefore, it was not possible to verify synergistic effects between glucoevatromonoside and acyclovir or glucoevatromonoside and furosemide. However, the inhibition caused by glucoevatromonoside on HSV protein levels could indicate that the Na+K+ATPase has been inhibited for this compound, as it is a cardenolide and its inhibition ability is well established. This inhibition could reduce the K+ concentration, and the HSV replication will not occur as usual.

Indeed, siApo-A1 treatment decreased

the cell proliferati

Indeed, siApo-A1 treatment decreased

the cell proliferation capacity of LoVo cells, although there was no significant Baf-A1 cost difference (Fig. 5B). Importantly, the level of c-PARP in normal cells under siApo-A1 exposure was clearly upregulated, suggesting that Apo-A1 acts as an apoptosis-preventing protein. Indeed, it was proposed that Apo-A1 might act as a regulator of tumor growth and metastasis [23]. However, considering that Apo-A1 is highly expressed in primary cancer cells rather than just in the secondary state [24], it is possible that this protein is involved in reversing malignant cells back into a normal cycle of differentiation. Recent findings that Apo-A1 is capable of promoting the cardiac differentiation of embryonic stem cells and inducing pluripotent stem cells [25] support this assumption. Therefore, our data and those of previous reports suggest that Apo-A1 is involved in the antiproliferative and proapoptotic activities of G-Rp1, via regulation of cancer cell differentiation. Relevant hypotheses regarding the JNK assay functional role of Apo-A1 in G-Rp1-mediated anticancer activity will be further tested in upcoming projects. In summary, we have demonstrated that G-Rp1 is capable of suppressing the proliferation of colorectal cancer cells and enhancing their apoptosis via enhanced levels

of Apo-A1. The protein levels of c-PARP and p53 were enhanced under siApo-A1 treatment, therefore, the Apo-A1-mediated anticancer effect of G-Rp1 might be linked to the functional involvement

of these proteins, as summarized mafosfamide in Fig. 6. Future studies will examine the exact molecular mechanism of Apo-A1-dependent G-Rp1 pharmacology in terms of its differentiation-inducing activities. The authors report no conflict of interests. “
“Alcoholism is a chronic relapsing disorder that is primarily driven by negative reinforcement via the reduction of withdrawal symptoms including anxiety, depression, hyperirritability, and insomnia. Of these symptoms, anxiety appears to be the most critical [1]. Abstinent alcoholics are more likely to return to drinking to ease psychological feelings of anxiety or depression, rather than to alleviate physical withdrawal symptoms. Similarly, ethanol-dependent rats exhibit elevated anxiety-like behaviors during ethanol withdrawal (EW) and excessive ethanol self-administration following a period of EW [2], and a number of pharmacological antianxiety agents reduce ethanol self-administration and the cue-induced reinstatement of alcohol seeking [3]. The central nucleus of the amygdala (CeA) is important for the integration of stress with the rewarding effects of ethanol and plays a crucial role in the development of anxiety and ethanol dependence [4].

65 ms, SD = 54 37 ms) compared to an easily discriminable pwin pa

65 ms, SD = 54.37 ms) compared to an easily discriminable pwin pair (80/20, mean = 959.67 ms, SD = 42.51 ms) (F[1, 15] = 125.81, p < 0.0001, η2 = 0.89). There was also a linear effect of test number with participants becoming quicker with time (F[1, 15] = 35.65, p < 0.0001, η2 = 0.70). There were no effects of session (mean actor = 1038.63 ms, SD actor = 51.01 ms; mean observer = 1066.69, SD observer = 49.67 ms), showing that any difference

found between observational and operant learning was not explicable by GS-7340 mouse RT differences. The results from Experiment 1 show that, while value learning through trial-and-error is highly accurate, observational learning is associated with erroneous learning of low-value options (i.e. those with the lowest probability of reward). In essence, observational learners show a striking over-estimation of the likelihood of winning from the lower-value options, a fallacy leading to impaired accuracy when choosing between two low-value options. This learning difference was apparent even though monetary incentives and visual information were matched in actor and observer learning. A different number of test trials were paid for observers relative to actors and this might have had a general

effect on performance. However, it cannot explain observers’ asymmetrically poor accuracy when learn more choosing between the 40/20 gamble pairs, and financial incentives were matched across each learning session overall. It is important to note that over-estimation of the value of the 20% win option did not cause observers to perform significantly worse when choosing between the 80/20 pairs. This is likely to reflect the fact that the probability difference is

uniquely high for such pairs, allowing for lower uncertainty when determining the higher value choice. It is interesting to observe that individual choice accuracies do not asymptote to 100%, as might be expected from rational decision makers once they accurately learn the value of stimuli. This may partially reflect the phenomenon of probability matching, a common finding in learning experiments (Herrnstein, 1961, Lau and Glimcher, 2005 and Sugrue et al., 2004), arising from a matching of choice frequency to average reinforcement rate. Note that, in our data, choice Prostatic acid phosphatase frequencies do not simply match learnt probabilities of reward, moreover probability matching does not in itself predict any difference between acting and observational learning. Two potential design weaknesses can be identified in Experiment 1. First, by yoking the sequence of actor choices to participants’ subsequent observer session, to match actor and observer learning for information presented, we are not able to counterbalance session order. Since participants also learnt about novel stimuli in the second session, learning may be worse solely because the task has switched.

, 2011) The dam-related processes have also altered the transpor

, 2011). The dam-related processes have also altered the transport of Huanghe material to the sea. The annual WSM scheme has imposed an extreme disturbance on the transport pattern of Huanghe organic carbon, silicon, and phosphorus (He et al., 2010). During the 2003–2009 WSM, large proportions of the annual dissolved organic carbon (35%) and particulate organic carbon

(56%) were transported to the sea. This dam-controlled input of organic carbon has a series of potential impacts on the biogeochemical processes at the river selleckchem mouth and its ambient sea (Zhang et al., 2013). Similarly for the Danube River, dissolved silicate load of the river had been reduced by about two thirds since dam constructions in early 1970s, which resulted in a series of environmental problems in the Black Sea (Humborg et al., 1997). The construction of Three Gorges Dam has potential impacts on the ecosystem in the Yangtze estuary and coastal waters where eutrophication and harmful algal bloom frequently occur.

The Yangtze River is estimated to lose a considerable proportion of its annual nutrient (in particular phosphorous and silicon) flux to the sea (Wang and Uwe, 2008), primarily due to dam-related processes. For the Mekong River, the trapping of nutrient-rich sediment by dams would potentially lead to decline in agricultural productivity and loss of agriculture land in the Mekong river delta. The damming of large rivers has therefore received both positive and negative feedbacks. learn more As stated by Milliman (1997), river damming is a double-edge sword. The four large dams on the Chinese Huanghe have altered its water and sediment fluxes, suspended sediment concentration, grain sizes, and inter-annual patterns of water and sediment delivery to the sea. In detail,

the dam effects on the Huanghe can be summarized as follows: (1) The four large new dams modulate the river flow between wet and dry seasons. Flow regulations lead to increases in water consumption over the watershed, a dominant cause for decreasing Huanghe material to the sea. Huanghe water discharge to the sea now relies heavily on Xiaolangdi releasing practices. Damming of the Huanghe has received both positive and negative feedbacks. Infilling of sediment behind the Xiaolangdi dam remains high and riverbed scouring began to weaken after 2006. It will be a big problem finding a location for the sediment when of the Xiaolangdi reservoir eventually loses its impoundment capacity. The Huanghe provides an example of management issues when large dams eventually lose their impoundment capacity. This study is jointly funded by the Youth Foundation of State Oceania Administration, China (No. 2010309) and the National Special Research Fund for Non-Profit Sector (No. 200805063 and No. 201205001). We gratefully appreciate the chief editor and the anonymous reviewers for their helpful comments which improved the manuscript.