Anti-obese and anti-diabetic effects of a mixture of daidzin and glycitin on C57BL/6J mice fed with a high-fat diet

Abstract

We investigated the effects of a mixture of daidzin and glycitin, which are the glycoside-form isoflavones of daidzein and glycitein, respectively, on body weight, lipid levels, diabetic markers, and metabolism in a high-fat diet (HF) fed C57BL/6J mice for 92 days. The mice were divided into basic diet group (CON), HF group, and HF companied with the isoflavone mixture group (HFISO). Results showed that mice in HFISO had a significantly lower body weight and adipose tissue compared to HF group. Blood glucose, serum HbA_1c, and serum insulin also showed lower levels in HFISO group. In addition, higher hepatic GSH level and lower serum 8-hydroxy-2′-deoxyguanosine (8-OHdG) level were found in HFISO group mice. This suggests that the regulation of oxidative stress by daidzin and glycitin was closely related to the suppression of adipose tissue and the progression of diabetes.

Graphical Abstract

Adipose cells enlargement occurs on high-fat diet feeding mice. The supplementary of isoflavones may regulate the oxidative stress, and suppress the adipose tissue and the progression of diabetes.

Key words:

• C57BL/6J mice
• daidzin and glycitin mixture
• obesity
• diabetes
• oxidative stress

The prevalence of obesity and its related diseases has been increasing worldwide, and it is considered to be a disorder of energy balance, and is associated with hyper insulinemia, insulin resistance, and abnormalities in lipid metabolism.¹⁾ It is one of the most important risk factors in the development of type 2 diabetes, atherosclerosis, and certain cancers.^1,2) Obesity has increased alarmingly since 90s, due to lifestyle risk including less physical activity and higher energy consumption.³⁾

Studies suggest that Asians, who consume 9–30 g soybeans per day,⁴⁾ experience lower incidences of some diseases, including prostate cancer, cardiovascular disease, osteoporosis, obesity, and related diseases.^5–7) Soybeans (Glycine max L. Merril) have long been an important protein source, complementing grain proteins. In addition to protein, soybeans also contain various nutrients and functional components.⁷⁾ The above health benefits associated with soy consumption have been linked to the content of isoflavones, the main class of the phytoestrogens.^2,5) Soybean isoflavones are structurally similar to estrogen, bind to estrogen receptors, and exhibit weak estrogenic activity.⁶⁾ It has been reported that isoflavones play an important role in the prevention of hormone-dependent diseases, osteoporosis, cardiovascular diseases, cancer, and postmenopausal syndrome.^7–10)

In nature, isoflavones occur in more than 300 kinds of plants, mostly in roots and seeds.¹¹⁾ Isoflavones exist in both the aglycone and glycoside forms in soybean foods. Glycoside conjugates of isoflavones are the major naturally occurring isoflavones in soybean and soybean-based food products.¹²⁾ On average, cooked soybeans, texturized vegetable protein, and soy milk powder contain >95% of the total isoflavones as glucoside.¹²⁾ So, it is meaningful to study the bioactivity of the glucoside-form isoflavones. Some studies have proved that the aglycone forms are absorbed more efficiently than their glycosides and superior in preventing some diseases.^13,14) Other study reported that the bioavailability of glycoside-form isoflavones is greater than their aglycone form.¹⁵⁾ Study has also shown no difference in bioavailability between glycoside-form isoflavones and their aglycones.¹⁶⁾

Health effects of daidzin have been proved on memory impairment, hepatic failure, and glucose homeostasis.^17–19) It was also reported that glycitin has preventive effect on bone loss and osteonecrosis.^20,21) However, to our knowledge, there are few studies that have been done on the effects of glycoside-form isoflavone supplementation on lipid metabolism associated with diabetes mellitus. In the present study, we have studied the changes caused by both diet and treatment with a mixture of daidzin and glycitin which are the glycoside forms of daidzein and glycitein, respectively, on lipid and glucose metabolism, in order to understand the mechanism of action of isoflavones in improving metabolic parameters. Additionally, we investigated the antioxidant levels which may have the beneficial effects on the blood glucose profiles in high-fat diet-induced obese mice.

The daidzin and glycitin mixture was the by-product in the production process of soy isoflavone supplements. We hope our study will be helpful to the utilization of the pure glycoside-form isoflavone mixture.

Materials and methods

Animals and diet

Male C57BL/6J mice with an average weight of 19 g (6 weeks old) were purchased from CLEA Japan Inc. (Tokyo, Japan) and housed individually under a 12:12 h light–dark cycle at 22 ± 2 °C and 40–60% humidity. After acclimatizing for 5 days, the C57BL/6J mice were randomly divided into 3 groups of 7 each, which fed on either the basic diet (CON group), or high-fat diet (HF group), or on the high-fat diet containing the isoflavone mixture (daizin and glycitin) at 0.06% (HFISO group). The composition of each experimental diet is shown in Table 1. The isoflavone mixture provided by Fuji Oil Co. Ltd (Osaka, Japan) was composed of daidzin and glycitin with a ratio of about 3:2. The diets and water were given for 92 days ad libitum. The body weight was measured every two days and the food intake was measured daily. At the end of the feeding period, the mice were anesthetized with Nembutal (Dainippon Pharmaceutical Co., Osaka, Japan) after 10 h of fasting, and the blood was collected by a cardiac puncture, followed by detaching the liver and stored at −80 °C until needed for analysis. The left perirenal adipose tissues, left epididymal adipose tissue, and mesenteric adipose tissue were also detached and stored at −80 °C after weighing. The serum was prepared by centrifuging the blood at 3000 ×g for 10 min. Liver lipids were extracted by the method of Floch.²²⁾

Table 1. Composition of the diet (%).

Constituent	CON	HF	HFISO
Casein	20	20	19.94
α-Cornstarch	53	8	8
Sucrose	10	25	25
Cellulose powder	5	5	5
Corn oil	7	7	7
Tallow	0	30	30
Mineral mixture^a	3.5	3.5	3.5
Vitamin mixture^b	1	1	1
L-Cystine	0.3	0.3	0.3
Choline bitartrate	0.2	0.2	0.2
Diadzin and glycitin mixture^c			0.06

^a AIN-93G-MX.

^b AIN-93-VX were obtained from Oriental Yeast Co. Ltd.

^c Diadzin and glycitin mixture (obtained from Fuji Oil Co. Ltd (Osaka, Japan) contains 97% of total pure isoflavone with a ratio of about daidzin:glycitin = 3:2).

The mice were cared for according to the institutional guidelines of Yamagata University Japan.

Fasting blood glucose and HbA_1c levels

Fasting blood glucose levels were measured with a Medesafe GR-102, 20–600 mg/dL measuring range (Termo Co., Tokyo, Japan) 10 h after fasting on day 1, 42, and 92 of the feeding period. HbA_1c levels were measured with a Micromat II (Bio-Rad Laboratories, California, USA) at 10:00–11:00 am after 10 h of fasting on day 42 and 92 of the feeding period.

Lipid analysis

Total cholesterol (TC) and high-density lipoprotein cholesterol (HDL-Chol), triglyceride (TG), phospholipid (PL), and total non-esterified fatty acids (NEFA) in the serum and liver, and feces TC, TG, and total bile acid levels were measured with commercial kits (cholesterol E-test, HDL-cholesterol E-test, triglyceride E-test, phospholipid B-test, NEFA C-test, and total bile acid test; Wako Pure Chemical Industries, Osaka, Japan). Liver and feces lipids were measured by using the isopropyl alcohol-soluble fraction of the lipids, which was prepared by removing the solvent in the lipid fraction obtained by the method of Folch and Lees.²²⁾

Measurements of the insulin and the other cytokines

The serum levels of insulin, adiponectin, and high molecular weight (HMW) adiponectin were measured with commercial ELISA kits (each, Mouse insulin (S-Type) Kit, Shibayagi Co., Gunma, Japan; Mouse adiponectin kit, Otsuka Co., Tokyo, Japan; Mouse high molecular weight adiponectin kit, Shibayagi Co., Gunma, Japan; Mouse leptin Kit, Morinaga Co., Tokyo, Japan).

Measurements of GSH and 8-OGdG

For the measurement of liver GSH, 50 mg of liver was homogenized with 500 μL of MeOH containing each of (20 μM of) methionine sulfone, MES, and D-Camphor-10-sulfonic acid, followed by the addition of 500 μL of chloroform and 200μL of Milli-Q water and by mixing. Subsequently, the mixture was centrifuged at 4600 ×g at 4 °C for 15 min. The upper phase was ultra-filtered with an ultra centrifugal filter (Millipore 5-KDa cutoff filter; Millipore Co., Japan) at 9000 ×g for 3.5 h (at 4 °C). The obtained filtrate was concentrated to dryness by centrifugation at 35 °C for 3 h, followed by solublization with Milli-Q water containing 3-aminopyrrolidine and trimesate (each, 200 μM), and analyzed by CE-TOFMS (capillary electrophoresis time-of–flight mass spectrometry).²³⁾

The serum 8-OHdG was measured by using an ELISA kit, 8-OHdG Check (Japan institute for the control of Aging, NIKKENSEIL Co. Ltd, Fukuroi–shi, Shizuoka Prefecture, Japan).

Measurements of liver enzyme activities

A sample to measure the liver enzyme activities was prepared by homogenizing the liver in a 3 mM tris–HCl buffer (pH 7.2) containing 0.25 M sucrose and 1 mM EDTA. The supernatant of the homogenate obtained by centrifuging at 500 ×g for 10 min at 4 °C was re-centrifuged at 9000 ×g for 10 min at 4 °C, and further centrifuged at 105,000 ×g for 60 min.

Fatty acid synthesis (FAS) activity was determined in terms of malonyl-CoA and acetyl-CoA-dependent oxidation of NADPH following the methodology of Kumar and Carey.^24,25) The reaction mixture was composed of a 0.1 M phosphate buffer (PH 7.0) containing 0.2 mM EDTA, 0.3 mM NADPH, 0.05 mM acetyl-CoA, 0.2 mM malonyl-CoA, and the sample solution. The rate of decrease in the absorbance at 340 nm was measured.

The reaction mixture for measuring the carnitine palmitoyl transferase (CPT) activity was composed of a 58 mM tris–HCl buffer (pH 8.0) containing 0.25 mM DTNB, 0.04 mM palmitoyl-CoA, 1.25 mM EDTA, and 1.25 mM L-carnitine. The CPT activity was determined from the rate of change in absorbance at 412 nm.²⁶⁾

Fecal lipid analyses

Feces were collected on days 86–88 for 72 h. Total fecal lipid was extracted by the method of Folch and Lees.²²⁾ The lipids in the extracts were measured with commercial kits as described above. Total bile acid (T-BA) levels were measured by the Total bile acid test kit (Wako pure Chemical Industries, Osaka, Japan).

Statistical analysis

Each value was given as the mean ± SEM. The homogeneity of variance between treatments was verified by Bartlett’s test. Data were statistically analyzed by a one-way analysis of variance. A post hoc analysis of significance was made by using Fisher’s PLSD test, where differences were considered significant at p < 0.05.

Results

Food intake, body weight, and organ weights

No differences were found between the HF and the HFISO groups on food intake; body weight gain in the HFISO group was lower than that in the HF group, but no statistical difference was found (Table 2). Body weight after 4 weeks in the HFISO group showed lower levels than those in the HF group (Fig. 1).

Table 2. Effects of isoflavone mixture (daidzin + glycitin) on the metabolic parameters.

Measurements	CON	HF	HFISO
Initial body weight (g/mouse)	19.4 ± 0.32^a	19.3 ± 0.31^a	19.7 ± 0.58^a
Food intake (g/d/mouse)	4.95 ± 0.11^a	4.61 ± 0.21^a^b	4.29 ± 0.11^b
Body weight gain (g/92 days)	10.1 ± 0.4^b	16.8 ± 1.3^a	13.8 ± 1.3^a
Liver weight (% of body weight)	3.54 ± 0.05^a	3.38 ± 0.06^a	3.61 ± 0.09^a
Kidney weight (% of body weight)	0.65 ± 0.02^a	0.58 ± 0.03^a	0.64 ± 0.03^a
Serum lipid levels
T-Chol (mg/dL)	106 ± 3^a	111 ± 7^a	105 ± 3^a
TG (mg/dL)	50.8 ± 3.5^a	38.1 ± 3.5^a	39.3 ± 4.1^a
NEFA (mEq/dL)	0.69 ± 0.03^a	0.72 ± 0.03^a	0.67 ± 0.02^a
HDL-Chol (mg/dL)	91.0 ± 9.0^a^b	68.9 ± 8.0^b	98.9 ± 8.8^a
Atherogenic index^1	0.30 ± 0.19^b	0.90 ± 0.23^a	0.21 ± 0.09^b
Serum adiponectin (ug/mL)	2.02 ± 0.08^a	1.91 ± 0.10^a	1.91 ± 0.06^a
Serum leptin (ng/mL)	13.3 ± 4.6^b	93.2 ± 28.2^a	30.1 ± 10.9^b
Serum TNF-α (pg/mL)	181 ± 7^a	252 ± 75^a	174 ± 16^a
Liver lipid levels
TC (mg/g of liver)	2.71 ± 0.24^a	2.48 ± 0.20^a	2.72 ± 0.18^a
TG (mg/g of liver)	42.3 ± 2.9^a	35.5 ± 2.9^a	35.6 ± 2.8^a
PL (mg/g of liver)	16.5 ± 0.6^a	13.8 ± 1.0^b	15.5 ± 0.6^b
NEFA (mEq/dL)	8.40 ± 0.90^b	8.11 ± 0.92^b	13.1 ± 0.41^a
Liver FAS activity (μmol/ min/mg protein)	1.93 ± 0.20^a	1.00 ± 0.13^b	1.22 ± 0.14^b
Liver CPT activity (μmol/min/mg protein)	0.71 ± 0.04^a	0.67 ± 0.06^a	0.64 ± 0.06^a
Feces lipid levels
T-Chol (mg/3 days of feces)	5.92 ± 0.31^a	5.11 ± 0.59^a	5.00 ± 0.54^a
TG (mg/3 days of feces)	1.30 ± 0.19^b	3.44 ± 0.37^a	3.03 ± 0.24^a
Total bile acid (mg/3 days of feces)	0.89 ± 0.07^b	1.75 ± 0.21^a	1.45 ± 0.13^a

Notes:

T-Chol, total cholesterol; HDL-Chol, high-dennsity lipoprotein-associated cholesterol; TG, triglyceride; NEFA, non-esterified fatty acid.

¹ (Total-Chol−HDL-Chol)/HDL-Chol). Each value is mean ± SEM. n = 5–7 for each group. Values without a common letter differ (a, b) significantly (p < 0.05). Compare with HF group. *0.05 < p < 0.1.

Fig. 1. Effect of dietary isoflavone mixture (daidzin + glycitin) on the body weight.

Notes: Each value is expressed as the mean ± SEM. n = 6–7 for each group. Values without a common letter differ significantly (p < 0.05). Compare with HF group. *0.05 < p < 0.1.

Liver and kidney weights did not differ among the CON, HF, and HFISO groups. Perirenal and mesenteric adipose tissue weights also did not differ between the HF and HFISO groups, but the epididymal and visceral adipose tissue weights (the total weight of perirenal, epididymal, and mesenteric adipose tissues) in the HFISO group were significantly lower than those of the HF group (Fig. 2).

Fig. 2. Effect of dietary isoflavone mixture on the adipose tissue.

Notes: Each value is expressed as the mean ± SEM. n = 5–7 for each group. Values without a common letter differ significantly (p < 0.05). CON, basal diet group; HF, high-fat diet group; and HFISO, high-fat diet group with isoflavone mixture.

Fasting blood glucose, HbA_1c, and insulin levels

The fasting blood glucose levels on day 1 did not differ among CON, HF, and HFISO groups, but it tended to be lower on day 42, and significantly lower on day 92 in the HFISO group compared to the HF group (Fig. 3(A)). The serum HbA_1c level of the HFISO group showed a significantly lower level on day 42 and day 92 than that of the HF group (Fig. 3(B)). Blood insulin was also significantly decreased in HFISO group compared to the HF group (Fig. 3(C)). No difference was found among the groups in OGTT which was carried out on day 42 (Fig. 3(D)).

Fig. 3. Effect of dietary isoflavone mixture on the fasting blood glucose (A), serum HbA1_c (B), insulin (C), and OGTT (D) levels.

Notes: Each value is expressed as the mean ± SEM. n = 5–7 for each group. Values without a common letter differ significantly (p < 0.05). CON, basal diet group; HF, high-fat diet group; and HFISO, high-fat diet group with isoflavone mixture.

Serum lipid and cytokine levels

There were no statistical differences in the serum TC, TG, and NEFA levels among the CON, HF, and HFISO groups. The HDL-Chol level was significantly higher in the HFISO group, compared to the HF group. The Atherogenic index was significantly lower in the HFISO group than in the HF group. The adiponectin and HMW adiponectin levels did not differ among the three groups (Table 2), but serum insulin and leptin levels in HFISO group were significantly decreased compared to HF group (Fig. 3(C), Table 2).

Liver lipid levels and enzyme activities

There were no differences on liver lipid levels including TC, TG, and PL, or on the activities of liver FAS and CPT enzymes between HF and HFISO groups (Table 2).

Feces lipid levels

There were no differences in the fecal excretion of TC, TG, and T-BA between the HFISO and HF groups.

Serum 8-OHdG and liver GSH levels

As shown in the Fig. 4(A), the level of reduced-form GSH of the HF group was significantly lower than that of the CON and HFISO groups. In contrast to the GSH level, the oxidized-form glutathione (GSSG) level of the HF group showed a higher tendency than that of the HFISO group. The serum 8-OHdG level of the HFISO group was significantly lower than that of the HF group (Fig. 4(B)).

Fig. 4. Effects of the dietary isoflavone mixture on the liver glutathione (a) and the serum 8-hydroxy 2′-deoxyguanosine (b) levels.

Notes: CON, basal diet group; HF, high-fat diet group; HFISO, isoflavone mixture-added high-fat diet group. GSH, glutathione reduced form; and GSSG, glutathione oxidized form.

Discussion

Obesity is considered to be the cause of metabolic syndromes, because it causes the accumulation of visceral lipids, insulin resistance, increased serum low-density lipoprotein cholesterol (LDL-Chol) level, and decreased HDL-Chol level, which are all closely concerned with complications of metabolic syndrome.²⁾ Several authors reported benefits of isoflavone consumption on improvement of the serum TG and on the prevention of obesity and diabetes.^10,26) However, negative effects of isoflavones have also been reported that there are no significant effect on serum lipids,²⁷⁾ plasma TG, plasma glucose, or insulin concentrations,²⁸⁾ suggesting a necessity to determine the effects of isoflavones more precisely. Almost all of the experiments concerned with these reports were carried out using genistein and daidzein as aglycones, but determination of the physiological functions of daidzin and glycitin, which are isoflavone glycosides, has not yet been fully studied in C57BL/6J mice.

In the present study, body weight was significantly lower in the HFISO group than that in the HF group from week 4 of the feeding period (Fig. 2). This phenomenon resembled that in the previous reports, in which it was reported that daidzein causes depression in body weight in mice and rats,^29,30) suggesting that daidzin might have the effect of decreasing body weight gain as the glycoside-form daidzein. It is known that a high-fat diet increases adipose tissue, which is a crucial factor in the development of obesity.³¹⁾ Obesity is defined as excess adipose tissue that can produce adverse health effects.³²⁾ In this study, epididymal and visceral adipose tissue weight in the HFISO group decreased significantly (Fig. 3), showing a possibility that the isoflavone mixture had an activity to inhibit the accumulation of adipose tissue. Other experiments in rats have also demonstrated that dietary isoflavones significantly decrease body weight and adipose tissue.³³⁾

In the present study, it has been observed that the serum HDL-Chol level was increased significantly by feeding the isoflavones, similar results were also reported in the previous studies.^26,34) In contrast to HDL-Chol level, serum TC and TG levels did not differ by significant levels between the HF and HFISO groups. Similar study showed that soy isoflavone supplementation had no significant effect on the reduction of TG and TC levels in fasting plasma in diabetic rats.²⁸⁾ But some other studies reported different results. Lee demonstrated that the concentrations of serum TG and TC were significantly reduced by the administration of genistein.³⁵⁾ Sosic et al. reported that the serum TC was decreased while the serum TG increased by subcutaneous administration of genistein and daidzein in rats.³⁶⁾ The opinions about the effect of isoflavones on serum lipid files were divided. In this study, we observed that the effects of the glycoside-form isoflavone diadzin and glycitin on serum lipids were minimal.

We hypothesized that hepatic lipid content might be reduced by isoflavones supplementation in the present study. However, the concentrations of hepatic TG and TC were not significantly reduced in the group supplemented with isoflavone mixture compared to the control group. Previous study showed a significant reduction in hepatic fat contents in diabetic rats which were fed a diet supplemented with genistein at the level of 60 mg/100 g for 3 weeks.³⁵⁾ We administrated the mice with isoflavones at the same amount for 92 days. It is supposed that the effect of isoflavones on hepatic lipid may partly depend on the feeding period and the species of the animals.

Numerous studies have been done on the effect of isoflavone on improvement of diabetes. Zimmermann et al. reported that delaying the progression of diabetes in db/db mice was independent of the isoflavone concentration.³⁷⁾ Choi et al. reported that genistein and daidzein prevent diabetes in non-obese diabetic mice.³⁸⁾ In this study, fasting blood glucose and HbA_1c were both decreased in the administration of isoflavone on the feeding period of day 42 and day 92. A lower level in the serum insulin was also observed in HFISO group, but no difference was found in OGTT. Obesity is highly associated with insulin resistance and it is the biggest risk factor for non-insulin-dependent diabetes mellitus.³⁹⁾ The serum insulin level in this study was increased significantly in the HF group as it has been reported that feeding of a high-fat diet could increase the serum insulin level, inducing insulin resistance on C57BL/6J mice;³¹⁾ the lower insulin level in the HFISO group may indicate that the isoflavone mixture is available to improve insulin resistance which was caused by high-fat diet.

Leptin is a key component that regulates food intake and energy utilization, and thus plays a dominant role in ultimately regulating the size of the body’s energy stores.⁴⁰⁾ Leptin increased in diet-induced obese mice, accompanied with a failure of any decrease in food intake or body weight，this hyperleptinemia was viewed as evidence of leptin resisitance.⁴¹⁾ In this study, serum leptin level in HF group was significantly higher than that in CON group, indicating that leptin resistance was reduced by high-fat diet. The decreased leptin level in HFISO group indicates that the leptin resistance was improved by the supplementation of diadzin and glycitin, and further decreased the accumulation of body adipose tissue through regulating energy expenditure. The improvement of leptin resistance in the HFISO group was correlated with body weight and plasma insulin level.⁴²⁾

The free radical scavenging properties of isflavone are dependent on its polyphenolic structures, which function as donors of hydrogen atoms to deleterious oxy-radicals.⁴³⁾ In the present work, the level of hepatic reduced glutathione (GSH), a relevant antioxidant, was significantly increased, and the level of hepatic oxidized glutathione (GSSG), an oxidant compound, was decreased. Lower serum 8-OHdG was also observed, indicating that dietary of daidzin and glycitin mixture may improve the oxidative status in mice, and generation of ROS, which are concerned with production of 8-OHdG, is suppressed by feeding of the isoflavone mixture, and that ROS is efficiently scavenged by antioxidant enzymes necessary GSH in its action.

In summary, the present study has demonstrated that the dosage of 0.06% glycoside-form isoflavone mixture daidzin and glycitin is effective in decreasing body weight and adipose tissue in mice. The isoflavone mixture also decreases fasting serum glucose, HbA_1c, and insulin levels. The data suggest that daidzin and glycitin mixture may have favorable effects on the anti-obese and anti-diabetic in high-fat diet-induced C57BL/6J mice. Determining of the mechanism and which one is more effective between daidzin and glycitin requires further investigation.

Funding

Financial support for the present study was provided by the Scientific Research Foundation for Overseas scholars of Educational Commission of Heilongjiang Province of China [grant number 1253HQ0012]; the Initial Founding of Scientific Reasearch of Heilongjiang Bayi Agricultural University [grant number XYB2013-25]; and the National Natural Science Foundation [grant number 31171657].

Notes

Abbreviations: AUC, area under the curve; CPT, carnitine palmitoyl transferase; FAS, fatty acid synthase; HDL-Chol, high-density lipoprotein cholesterol; HbA1_c, hemoglobin A1_c; HMW adiponectin, high molecular weight adiponectin; NEFA, non-esterified fatty acids; OGTT, oral glucose tolerance test; T-BA, total bile acid; T-Chol, total-cholesterol; TG, triglyceride; TNF-α, tumor necrosis factor-α; PL, phospholipids; 8-OHdG, 8-hydroxy-2′-deoxyguanosine.

Author contribution: Animal experiments and manuscript preparation (YZ), providing sample and experimental places (KI), and data analysis and manuscript revision (CY).