カイコ体液中のキモトリプシンインヒビターの生 理・生化学的研究

九州大学学術情報リポジトリ

Kyushu University Institutional Repository

カイコ体液中のキモトリプシンインヒビターの生 理・生化学的研究

白井, 孝治

九州大学農学研究科遺伝子資源工学専攻

https://doi.org/10.11501/3130902

出版情報：Kyushu University, 1997, 博士（農学）, 課程博士 バージョン：

権利関係：

Chapter 3

Involvement of CI-8 in apoptosis of the midgut during the larval-pupal metamorphosis

CI-8 inhibits the proteolytic activity of digestive juice (see Chapter 1 ).

Also other high-molecular-weight serpin type of Cis, e.g., inhibitor-d (Eguchi and Shomoto, 1985) and sw-Achy (Sasaki and Suzuki, 1982; Sasaki and Ito,

1985),

have anti-proteolytic activities against the chymotrypsin-like digestive juice proteinases named 6B3 (Eguchi and Shomoto, 1985) and P-Ill (Sasaki

and Suzuki, 1982). The physiological significance of inhibitory activity of humoral Cis vs. digestive juice proteases was uncertain (supposedly involved in a self-defense system restraining the effects of digestive juice extruded into the hemolymph via wounded larval midgut cells; Eguchi ^et

al.,

1979a).

The midgut is an important organ for digestion and assimilation of nutritional materials during feeding periods. After the spinning stage, the digestive system in the midgut becomes needless and the organ starts remodeling including histogenesis. Midgut tissues fall and migrate into the alimentary canal. This process is considered to be a kind of apoptosis, since during the larval-pupal metamorphosis DNA in the midgut receives nick and

shows a ladder pattern upon gel electrophoresis (Ishihara et al., 1994 ).

Similar phenomenon occurs in the degenerating silk glands (Takenaka ^et

al., 1995).

In mammals, hemocytes prey and digest apoptotic tissues, but during metamorphosis of B. mori, endogenous and/or digestive juice enzymes may attack the larval midgut cells. The degradation products will be utilized to the adult development.

96

This chapter deals with the trial to reveal the role of Cis in apoptosis of the larval midgut by examining the relationships of CI-8 and a digestive juice protease. Prior to these, a novel protease inhibited by CI-8 was purified from digestive juice. Also the presence of sequestered CI-8 in the midgut and other tissues was demonstrated by biochemical and histochemical methods using the CI-8 antiserum.

Materials and Methods

Materials

DEAE-Sephacel and Toyopearl HW55 were purchased from Pharmacia LKB Biotechnology Co. Ltd. and Tosoh c:o. Ltd., respectively. For Butyl­

Toyopearl, see Chapter 1. Commercially available protease inhibitors used (specified in Results) were from Sigtna Chern. Co. Ltd.

Animals

The cz strain was thoroughly used.

Preparation of larval digestive juice

After starving for 6 hr, larvae were kept in chloroform gas and allowed

to vomit digestive juice into an ice cold vessel. The digestive juice collected was briefly centrifuged to discard large pieces of mulberry leaves and stored at -40 OC until used.

Standard assay method of proteolytic activity

The reaction mixture ( ^1.55 ml) containing 0.32% casein, 0.065 mM

glycine-NaOH buffer, pH 11.0, and enzyme was incubated at 30°C for 2 hr.

The reaction was stopped by adding TCA solution (final 2%) and the mixture was left at room temperature for 30 min and centrifuged at 3,000 x g for 30 min. The supernatant was measured for the absorbance of 280 nm by a Hitachi spectrophotometer. Values of appropriate controls whose reaction was stopped immediately after addition of casein were subtracted. One unit of proteolytic activity was defined as the amount which gives the absorbance of 0.01.

Zymograms

Zymogram of proteolytic activity after native PAGE or SDS-PAGE with

casein or synthetic substrate. In the case of SDS-PAGE, the gels were soaked in 1% Triton X-100 at room temperature for 30 min before carried to the next step.

The gels were incubated

in 4o/t)

asein solution (diluted with 0.1 M glycine-NaOH buffer, pH 11) at roon1 tern perature for 1.5 hr and at 30°C for 2 hr, stained with Coomassie Brilliant Blue R250 and washed in

methanol/acetic acid/water

(4: 1

^: 5, v/v/v). Protease activity was observed as an unstained band on blue background.

Alternatively, the gels were incubated in the staining solution (10 mg/5 ml of N-acetyl-D, L -phenylalanine-

�

-naphthylester dissolved in N,N'­

dimethylformamide plus 50 mg/50 ml of tetrazotized orthodianisidine dissolved in sodium phosphate buffer, pH 8) at 30°C for15 min. Protease activity was observed as a red band.

Zymogram of CI activity after SDS-PAGE

After SDS-PAGE, the gels were soaked in 1% Triton X-100 at room temperature for 30 min, and in a-chymotrypsin solution at room temperature for 10 min and then treated at 30°C for 10 min with N-acetyl-D, L-

98

phenylalanine-

�

-naphthylester plus tetrazotized orthodianisidine under the conditions as specified in Chapter 1 for zymogram after native PAGE.

Purification of a digestive protease

Digestive juice was fractionated by salting out with ammonium sulfate.

The precipitate between 20% and 80o/o saturation was collected by

centrifugation at 10,000 _x g for 30 min and dissolved in 0.01 M Tris-HCl buffer, pH 9.0, and extensively dialyzed against the same buffer. After

dialysis, the sample was fractionated by gel filtration using Toyopearl HW 55 equilibrated with the same buffer. The fractions containing protease were collected and precipitated by the addition of atnmonium sulfate up to 80%

saturation. The solution was dialyzed against 0.01 M Tris-HCl buffer, pH 9.0, containing 10o/o saturation of ammonium sulfate. The sample was subjected onto a Butyl-Toyopearl column equilibrated with the same buffer plus salt.

The column was washed with the same buffer plus salt, and then a linear

gradient was applied from 10 to 0% saturation of atnmonium sulfate in 0.01 M Tris -HCl buffer, pH 9.0. Active fractions were concentrated by Centriplus (Grace Japan

K. K.)

^and re-chromatographed with the same column

equilibrated with the same buffer. After elution, the fractions of protease were pooled and concentrated by Centricon, dialyzed against 0.01 M Tris -HCl buffer, pH 9.0, and then loaded onto a DEAE-Sephacel column. After the pass-through fraction was eluted completely, a linear gradient was applied from 0 to 0.3 M NaCl. All procedures except for ammonium sulfate precipitation were performed at 4 °C.

Determination of optimum pH and thermostability of protease

Optimum pH of protease was examined by the standard assay procedure except that pH was varied using the following buffers: sodium-phosphate

buffer for pHs 6 to 9, Tris-HCI buffer for pHs 7 to 9, glycine-NaOH buffer for pHs 9 to

11

and phosphate-NaOH buffer for pH

11

_to

12.

Thermostability of protease was tested after kept the enzyme solution

[ 150

units/ml of

0.01

M Tris-HCI buffer, pH 6.8, containing

10%

_(v/v)

glycerol] was kept at various temperatures from 4 to

1

_{OOOC for}

10

_{min. The}

proteolytic activity was then assayed by the standard procedure. Relative activities to that of the enzyme kept at OOC were calculated.

Effects of protease inhibitors and metal ions on activity of protease

The commercially available synthetic protease inhibitors used were specified in Results. The reaction mixture \Vith inhibitor at three

concentrations

(0.5, 1

_and

10

mM) was preincubated at

30oC

_for

20

_{min and}

then the activity was measured by the standard procedure.

Effects of various metallic ions on the activity of protease were studied by the standard procedure except that the reaction mixture contained various concentrations of Mg2+, Mn2+, Cu2+ or Ca2+ (all were chloride salts).

Preparation of digestive juice, midgut, midgut membrane, meconium and other tissue specimens

Digestive juice was collected into plastic tubes using a syringe. Also midgut was taken out, washed with physiological saline and, if necessary, separated into the three subfractions according to the regions anterior, middle and posterior (abbreviated as MG-a, MG-m and MG-p, respectively).

Peritrophic membrane was also prepared. Testes and ovaries were dissected from male and female larvae, respectively.

The cell membranes were prepared from midgut by the method described for the fat body in Chapter

2.

The tissues or membranes were each extracted by the procedure 100

described in Materials and Methods of Chapter 2 (under the subtitle

"Preparation of tissue extracts").

When subcellular fractions were needed, the middle plus posterior

regions of midgut were homogenized in three volutnes (v/w) of a fractionation buffer (3 mM Tris-HCl, pH 7.4, containing 0.25 M sucrose, 0.1 mM EDTA, 1 mM PMSF) and the homogenate was centrifuged at 700 x g for 10 min. After removing the fat layer, the supernatant was centrifuged again at 7,000 x g for

10 min. The precipitate was collected as P2. The supernatant was centrifuged at 105,000 x g for 60 min. The final supernatant and the pellet were collected as sup and P3, respectively. The P2 and P3 fractions were extracted and the supernatants were saved as described above.

Meconium was collected and centrifuged at 10,000 x g for 5 min to remove insoluble materials. The supernatant was saved for analysis.

Preparation of antisera

The antiserum against protease was prepared by injection of the purified protease with Freund's complete adjuvant into the abdominal cavity of a mouse four times at one week intervals. Antiserum for SP2 was kindly provided by Dr. Nagata at University of Tokyo.

Histochemical localization of proteolytic activity for artificial substrate in midgut sections and of localization of CI -8

Cryostat sections of midgut (semi whole-mount specimens) were

prepared by the same method as that described for fat body in Chapter 2. The frozen section on the slide grass was soaked in ice-cold absolute acetone for 30 min to fix the tissues and to remove lipid, reacted at room temperature for 30 min in the staining solution ( 10 mg/5 ml of N -acetyl-D, L -phenylalanine-b­

naphthyl ester dissolved inN ,N' -dimethylformamide plus 50 mg/50 ml of

tetrazotized orthodianisidine dissolved in sodium phosphate buffer, pH 8.0) and washed thoroughly in distilled water to stop the reaction. The proteolytic activity was visualized by red color.

Also the cryostat sections of the midgut were incubated with anti-CI-8 and anti SP2 serum and stained as described in Chapter 2.

Other procedures

Native PAGE, SDS-PAGE, isoelectric focusing, ELISA, Western blotting and other procedures including the analysis of oligosaccharide chain using PVDF membranes were performed as described in Chapters 1 and 2.

35S-labeled CI-8S was prepared as described in Chapter 2. Detection of the

receptor for CI -8 on the membrane fractions prepared from the midgut

specimens was carried out by the same method as that described for fat body in Chapter 2.

Results

Purification of a protease from digestive juice

Digestive juice from larvae on day 3 or 4 of the 5th instar of cz strain (after starving for 6 hr) was subjected to ammoni urn sulfate fractionation and then to gel filtration using Toyopearl HW 55. The protease activity was separated into two peaks (Fig. 3-1 ). The second peak (containing the fraction later named 35k protease) was further loaded onto a Butyl-Toyopearl column.

A linear gradient elution from 10 to 0% saturation of ammonium sulfate brought about five protease peaks (Fig. 3-2), of which the third one was re­

chromatographed with the same column (Fig. 3-3). The protease fractions

102

2.0

•

*

-0 (X) C\1

-( -

c: 1.0

"G)

� 0 _....

a.

0.0

0 20

35k protease

Protein Activity

40

� ....

60 fraction No.

80 100

2.0

-0 (X) C\1

-( -

>.

1.0 � ·:;;

; 0 cu (I) en (I) cu

� 0 _....

a.

0.0

Fig. 3-1. Elution profile of proteolytic activity of digestive juice.

Gel filtration of ammonium sulfate fraction

(20-80%)

^of

digestive juice protease on a Toyopearl HW55 column

(2.8

^x

70

em), equilibrated with

0.01

^M Tris-HCI buffer, pH

9.0.

^Fractions

of 5 ml were collected.

2.0 . ^{' \ ..} _{.. ..}

'

-0 CX) C\1

<t

-_c: 1.0

"G) ...

0 ...

0..

0.0

0 20

.. .. .. ' _\

.. . .. ' .. .. .. ' ' ..

' _\ .. . .. ' ' .. ..

' ..

• Protein .. Activity

40 Fraction No.

0 1.0

0.8 _-

0 co

7.5 _-^C\1 <t

.. _\

0.6 ... ^>-

·:;:

:;; 0 co

0.4 _tn ^(I)

15 ... co ^(I)

0 ...

0.2 ^0..

60

Fig.

3-2.

Elution profile of digestive protease on a Butyl-Toyopearl column.

Elution profile of digestive protease on a Butyl-Toyopearl column

(2.4

^x

20 em), equilibrated with 0.0 1 M Tris-HCl buffer, pH 9, containing 15%

saturation of ammonium sulfate. The absorbed materials were eluted with a linear gradient of ammonium sulfate from 15% to 0% saturation. Each fraction contained 5 mi.

104

0.30

-0 0.20

co C\1

<

-

"G) ... c:

0 a.. 0.10

0..

0.00

• Protein -· Activity 1.0

0.8

0.6

0.4

0.2

�� ... --�-r--_---,�---,---r---r--+ 0.0

0 10 20

Fraction No.

30 40

-0 CX) C\1

<

-

>.

:!::::

>

;::

0 m C1) 0 m C1)

...

0 a..

0..

Fig. 3-3. Rechromatography of 35k protease on Butyl-Toyopearl.

The column was equilibrated and eluted under the same conditions as those described in Fig. 3-2.

obtained were collected and loaded onto a DEAE-Sephacel column, which was eluted with a linear gradient from 0 to 0.3 M NaCl (Fig. 3-4). The first peak was the passed-thorough fraction and the second peak was saved as the final preparation. Typical purification procedures are summarized in Table 3-1.

The purification ratio of this protease from hemolymph was about 30-fold with a recovery of 0.08%.

-0 (X) C\1

<t

-

·a; c:

... 0 ...

Q.

0.08 • _Protein 2.0

.. Activity 35k protease

0.06 _... 1.5

0.04 1.0

0.02 0.5

0.00

4---r-�----r---�--y--�---,r-�-+

^0.0

0 10 20 30 40 50

Fraction No.

-0 (X) C\1

<t

-

>- ::: >

;:; 0 cu C1) 0 cu C1) ... 0 ...

Q.

Fig. 3-4. Ion-exchange chromatography of digestive protease on a DEAE-Sephacel column.

Ion-exchange chromatography of digestive protease on a DEAE-Sephacel column (0.8 ^x 10 em), equilibrated with 0.01 M Tris-HCl buffer, pH 9.0, and eluted with a linear gradient of NaCl from 0 to 0.3 M. Fractions of 5 ml were collected. The fractions containing 35k protease were indicated by an horizontal arrow.

Table. 3-1. Summary of purification of 35k protease.

Volume of starting digestive juice

(OJ)

was 50 ml. The total protein contents and the total activities were assayed after collecting the active fractions at each steps during purification procedure.

Total Total Specific Recovery protein activity activity

(o/o)

^Yield

(mg) (units) (units/mg)

OJ 660 27000 40.8 100 1.0

(NH4)2S04 63 6100 96.8 22.6 2.37

Toyopearl HW55 26 3350 129.0 12.4 3.15 Butyl-Toyopearl 3.7 728 197.0 2.7 4.81 DEAE-Sephacel 0.018 22.2 1387.5 0.08 33.9

Catalytic property, molecular weight and pi of the purified protease

The final preparation gave one protein band on native PAGE and the activity to the synthetic substrate (see Materials and Methods) was detected at the same position of the gel (Fig. 3-SA). SDS-PAGE (Fig. 3.5B) gave a single band of protein. These results indicated that this enzyme was fairly pure. The synthetic substrate used was for chymotrypsin, and the present enzyme was, therefore, a hydrolase attacking the Phe-X peptide bond (carboxyl side). The molecular weight was about 35,000 as calibrated by the mobilities of L.M.M.

on SDS-PAGE (Fig. 3.5B), and the enzyme was named 35k protease. The pi

of 35k protease was determined to be 5.1 by gel isoelectric focusing (patterns not shown).

protein activity

35k protease

94,000- 67,000- 42,000-

29,000-

20,100-

14,400-

protein activity

L.M.M. 35k protease DJ

Fig. 3-5. Analyses of 35k protease by PAGE.

(A) Native PAGE (10%) followed by protein staining with Coomassie Brilliant Blue (left) and by zymogram for activity using the synthetic

substrate (right). (B) SDS-PAGE (15%) followed by protein staining with Coomassie Brilliant Blue (left) and by zymogram for activity using casein as a substrate (right). L.M.M. and DJ indicate molecular weight markers and digestive juice (on day 3 of the 5th instar), respectively. Arrows indicate the position of 35k protease.

108

Additional properties of 35k protease

ConA blotting analysis after SDS-PAGE exhibited a positive reaction (Fig. 3-6), indicating that 35k protease has a oligosaccharide chain(s) and would be aN-glycoprotein. In addition, several bands were detected in the lane of digestive juice and there may be several glycoproteins.

35k pro DJ

Fig. 3-6. Reaction of 35k protease with ConA.

After SDS-PAGE

(10 %),

the gel was blotted to a PVDF membrane, which was then reacted with ConA-peroxidase. 35k pro and DJ indicate purified 35k protease and digestive juice (days 3 of the 5th instar), respectively. Arrow shows the position of 35k protease.

The proteolytic activity of 35k protease did not change largely after heating for 10 min at various temperatures up to 60°C (Fig. 3-7). However, it dropped markedly at 65°C and was almost lost at 70°C. The overall

thermostability pattern of this protease was similar to those of Cls-3 and 8 (Chapter 1 ).

0 20 40 60 80 100

Temperature

Fig. 3-7. Thermostability of 35k protease.

Thermostability of 35k protease, which was treated at an indicated temperature for 10 min at pH 6.8 and measured for activity. The results were expressed in relative activity to that treated at 30°C.

The caseinolytic activity examined at various pHs from 6 to 12 (Fig 3-8) indicated the highest values at pHs ll to 12. Little activity was detected at pH 6, indicating that this enzyme is adapted to the alkaline conditions of digestive JUICe.

110

-� 0 -

>a .:!::::

>

; 0

� C1) >

; �

Ci) a:

100

50

0

5 6 7 8

e Na-phosphate buffer

--["""_j-- Tris-HCI buffer

Glycine-NaOH buffer --c>- Phosphate-NaOH buffer

9 10 11 12 13 pH

Fig. 3-8. Optimum pH of 35k protease.

Activity was measured at an indicated pH at 30°C.

Results were expressed in relative activity to that measured in glycine-NaOH buffer, pH 11.

Of the commercially available protease inhibitors tested, DFP

(diisopropyl fluorophosphate) and PMSF (phenylmethylenesulfonyl fluoride) showed the strongest inhibitory effects on the caseinolytic activity of 35k protease (Fig. 3-9). Antipain gave medium results. Pepstatin, leupeptin, IAA (iodoacetoamide) and EDTA (2Na) had little or no inhibitory potency upon 35k protease. These results suggested that the present enzyme is a kind of serine protease like other previously reported digestive juice proteases (Sasaki and Suzuki, 1982).

Mn2+ added to the reaction mixture at concentrations up to 10 mM inhibited the caseinolytic activity of 35k protease by 70% (Fig. 3-1 0). Cu2+

ion scarcely influenced the activity. Mg2+ and Ca2+ augmented the activity to 150% at 10 mM.

100

-

�

0

-

::: �

> 50

.. (J cu cu :1

"C

·u; (1) a:

0

b��

,

^....

'

^..

^,

^{... _}

�. � ^�):

�� �� � ..,

Inhibitors

•

�

•

10mM 1 mM O.SmM

Fig. 3-9. Effects of various inhjbitors (see text for details) on activity of 35k protease.

Protease was incubated with a inhibitor in 0.1 M glycine-NaOH buffer, pH 11, for 20 min at 30°C and measured for activity. Results were expressed in relative activity to that without treatment.

11 2

200

0 ^{I I} 0

, D

•

<!;i; . •

'\,,,,·0 --^---_--...0

....^.... _{_} ca2+

___[] ^Mg2+

Mn2+

__() _ ^cu2+

I I I II I I .1

I I ¹ I I I I II 1

Concentration (mM)

I I I II I 10

Fig. 3-10. Effects of metal ions on activity of 35k protease.

Activity was measured under the presence of ion. Results were expressed in relative activity to that without ion.

Inhibitory activity of CI-8 against 35k protease

It was confirmed that CI-8 inhibited the present chymotrypsin type protease (Table 3-2). At pHs 11, 20 units of CI-8 completely inhibited 100 units of 35k protease.

Table. 3-2. Inhibitory activity of CI-8 against 35k protease.

The activity of 35k protease ( 100 units) was measured under the presence of the indicated units of CI-8. The results were expressed in relative activity to that without CI-8.

Phosphate-NaOH buffer pH 11 ( 40 units) Phosphate-NaOH buffer pH 11 (20 units) Phosphate-NaOH buffer pH 11 ( 4 units)

Detection of 35k protease in digestive juice

3.9°/o 0.0°/o 94.1°/o

Upon native PAGE followed by activity staining using the synthetic substrate for chymotrypsin, crude digestive juice specimens from male and female larvae on day 1 of the 5th instar gave several bands, of which dual bands marked by arrows in Fig. 3-11 A were the most intense. Both were reactive to the antiserum against a 35k protease raised in the present study as seen in the Western blot analysis of the gels (Fig. 3-11B). These closely migrating components were tentatively called isozymes A and B. The

electrophoretic patterns of isozymes A and 8 are schematically summarized in Fig. 3-12, which includes the zymogram pattern with the synthetic substrate and casein after SDS-PAGE (original data not shown). Most individuals of the cz strain had both isozymes but some were only with A; these were named the AB type and the AA type, respectively.

114

(A) (B)

male female

1 2

Fig. 3-11. Analyses by PAGE of protease activity in crude digestive juice taken from male and female larvae on day 3 of the 5th instar.

Two individuals of each sex, numbered 1 _and 2, were used and applied to Lanes 1 and 2, respectively. Arrows indicates the positions of 35k

protease and its counterpart (see text). (A) Native PAGE (10%) followed by zymogram for CI activity using casein as a substrate. (B) SDS-PAGE (15o/o) followed by Western blotting with the anti-35k protease serum.

Lanes 1, hemolymph having a active band above that of 35k protease; Lanes 2, hemolymph having 35k protease.

type

AA A8

A� --^-

A_.

�8

AA A8

- --

12.5°/o SDS-PAGE

�8

Fig. 3-12. Schematic diagram indicating the composition of 35k protease as analyzed by native and SDS-PAGE.

Purified 35k protease was of the A type and B type may be its allelic counterpart.

Detection and localization of 35k protease antigens in the midgut

The specimens of anterior, middle and posterior regions of the midgut (MG-a, MG-m, MG-p, respectively) and the peritrophic membranes (PM) of MG were prepared from larvae on day 3 of the 5th instar and subjected to native PAGE followed by the Western blot analysis using the anti-35k protease·

serum (Fig. 3-13, left panel). MG-1n exhibited a positive sign for 35k protease in the form of the AB type. Also PM as well as digestive juice gave very strong bands of the AB type. MG-a showed weak band at the position of 35k protease A. A slowly migrating band in the lane of MG-p (as well as of digestive juice used as a marker) may be non-specific. The control test with normal serum gave no signs (right panel).

1 1 6

_ __ ...,..

35kpro�

anti 35k protease normal serum

Fig. 3-13. Analysis of 35k protease-related proteins in the midgut.

MG-a, MG-m, MG-p (anterior, middle and posterior regions of the midgut, respectively), the peritoneal membranes (PM) of MG-(m+p) and digestive juice

(DJ)

were prepared from larvae on day 3 of the 5th instar, each extracted (as necessary) without further

subfractionation and subjected to native PAGE (7.4%) followed by Western blot analysis using the anti-35k protease serum. The amount of each sample applied was corrected to have the same protein content. Arrows with solid and broken lines indicates the positions of 35k protease and unidentified material, respectively.

The control test was done with normal serum.

The P2 (700 x g) and P3 ( 105,000 x g) subcellular fractions of MG­

(m+p) exhibited a marked sign for 35k protease whereas 105,000 x g sup did not (Fig. 3-14). Also P2 and P3 of MG-a exhibited a sign although weaker compared to MG-(m+p). The significance of other bands than 35k is unknown at present.

In any event, these findings clearly indicated that antigen for 35k­

protease was bound to MG-m in early 5th instar larvae. This protein (a

precursor to 35k protease?) is likely to be synthesized in the epithelium cells of MG-m and enter the peri trophic membranes, where it may be trapped until secreted into the lumen.

MG-a MG-(m+p)

P2 P3 sup P2 P3 sup

over 100k Da

52-53k Da 35k protease

Fig. 3-14. Analysis of 35k protease-related proteins in the midgut subcellular fractions.

The P2 (700 X g) and P3 (105,000Xg) subcellular

fractions of MG-a and MG-(tn+p) were extracted and subjected to native PAGE (7.4%) followed by Western blot analysis using the anti-35k protease serum. The amount of each sample

applied was corrected to have the same protein content. Arrows stand for the positions of 35k protease and unidentified

materials.

118

Developmental changes in total activity of protease and antigen for 35k protease in digestive juice

The caseinolytic activity of digestive juice measured in toto remained

relatively high during the 5th instar, making a peak

^on

da

y 4

(Fig. 3-15). It

dropped to vestigial revel until the onset of spinning, and was almost lost on day 2 of spinning.

X 100

.., 1.2 c

� 0

"5.

0.8

0

�

:!::::

c:

:::J

0.4

0.0

5-1 5-2 5-3 5-4 5-5 8-1 8-2 P-1 stage

Fig. 3-15. Developmental changes in caseinolytic activity of digestive juice.

5-X, S-X and

P-1

indicate day of the 5th ins tar, spinning stage and pupal stage, respectively.

Western blot analysis of protein using the antiserum against 35k protease revealed a single positive band in digestive juice from larvae on day 3 of the 5th instar, but no band in digestive juice from day

1

spinning larvae (Fig. 3-

16).

The hemolymph specimens analyzed as a negative controls at both stages

gave no stgns. Digestive juice from spinning larvae gave only a few protein bands when the gel was stained by Coomassie Brilliant Blue (patterns not shown). Thus the rapid drop of apparent proteolytic activity in the digestive juice at the onset of spinning may be due to the disappearance of proteins from digestive juice.

5-3 5=1

81 OJ 81 OJ

Fig. 3-16. Comparison of digestive juice banding pattern for 35k protease-related antigens.

Comparison of digestive juice banding pattern for 35k protease-related antigens between day 3 of the 5th instar (5-3) and day 1 of spinning (S-1) as examined by Western blotting analysis after native PAGE

(7.4%).

Also hemolymph specimens were tested as negative controls.

Developmental changes in proteolytic activity of the midgut

The total caseinolytic activity of the midgut extracts was high from the 4th molt to day 4 of the 5th instar (Fig. 3-17). The activity of MG-a began to decrease after day 5, while that of MG-(m+p) did after day 4 of the 5th instar.

Interestingly, the midgut retains marked proteolytic activity at the onset of spinning when digestive juice almost lost its protease activity.

MG-a was higher than MG-(m+p) in relative activity measured in toto (Fig. 3-17), a situation in contrast to the results of Western blotting for 35k protease (see Figs. 3-13 and 3-14), wherein MG-a was fainter than MG-(m+p).

These findings indicated that other proteases than 35k were present in MG-a.

X 100 1.5

Q) ::J 1.0

0 0 :;::;

C) E

0 ,...

(;) 0.5

== c::

::J

0.0

• MG-a --()- MG-(m+p)

4M 5-1 5-2 5-3 5�4 5-5 S-1 s-2 stage

Fig. 3-17. Developmental changes in caseinolytic activity of the midgut.

MG-a and MG-(m+p) were separately analyzed. 4M means the 4th molt. For other comments see the legend to Fig. 3-15.

Localization of proteolytic activity in the midgut using artificial substrate

The proteolytic activity was histochemically localized on the cryostat sections of the midgut using the synthetic substrate for chymotrypsin. At the onset of spinning when the epithelium cells at the middle region of midgut has not yet been replaced by the newly reconstructed epithelium cells, the strong activity appeared as a red pigment formed by hydrolyzed substrate at

microvilli of the columnar cells (Fig. 3-18A). On day 2 of spinning, it was also seen in the cells but the intensity was very low (Fig. 3-18B). The newly constructed pupal midgut cells and basement membranes, already observable on day 2 (Fig. 3-18B), were scarcely stained, indicating that no active protease was present in the newly differentiated cells.

Effects of incubation at room temperature on proteolytic activity of digestive JUICe

When digestive juice taken out of larvae on day 1 of the 5th instar was kept at room temperature (26.8°C) for 24 hr, the activity was not affected giving the values as high at that of the control kept at OOC (Fig. 3-19A). The staining pattern on zymogram after native PAGE scarcely changed after the incubation (Fig. 3-19B).

Detection of antigens related to 35k protease and to CI-8 in the posterior part of midgut during metamorphosis

The midgut from day 2 of spinning to day 1 of pupal stage turns into black at the posterior region (MG-p, shaded region in Fig. 3-20A). It was this region where the larval epithelial cell layer was degraded and cast off into the lumen. When the extract of MG-p prepared from day 2 spinning larvae was subjected to SDS-PAGE followed by Western blotting separately using the antisera for 35k protease, CI -8 and SP2, very strong bands were detected for

(A)

Bm

^../'

(B)

Bm

p

(A)

(B)

...,..,.. -

�����·',...

� IMl

.,.,:.L·'

l�Bm

M

"LL. 1.

"LL.

"LL 1L

I I

�

.. , _I ::::: �"

lr.!I.,...IL:.

' 1 1\�r.

, .... _

�Bm-.

"Lf7L.7. :-..

'ZLLL/ '� �

••�9�9�•-��.ra---,s�-�--�--��---�-•�--�ar..��z�----�z�.���==�;�; .. . ;:

Fig. 3-18. Histochemical localization of proteolytic activity in the midgut from a larva on day 1 (A) and on day 2 (B) of spinning.

Left: the larva was cut to prepare semi whole-mount sections which were stained using the synthetic substrate. Right:

the results are schematically drawn. lMG, pMG, Bm and M

M N

r--

(A) (B)

_o' ^{e� �}

�'4$ ��

cP _e'b'

30 -s

Q. •

E

CD 20

.,

.._ 0

=l.

�

= _c: 10

:::l

0'

control treatment

Fig.

3-19.

Effects of incubation of crude digestive juice at room temperature.

Effects of incubation of crude digestive juice at room temperature

(26.8oC)

_for

24

hr on caseinolytic activity (A). Larvae on day

1

_of

the 5th ins tar were used. The control specimen was kept at

4 oC

_for

24

hr. Also the treated specimen as well as the control was subjected to native PAGE

(10%)

followed by zymogram analysis for

caseinolytic activity

(B).

124

(A)

(B)

68--69kD�

antl35k pro anti Cl-8 anti SP-2

Fig. 3-20. Detection of

^35k

protease-related and CI-8-related antigens in the posterior part of the midgut.

(A) Schematic drawing of the midgut during larval-pupal

metamorphosis. (B) The shaded area was prepared from day 1

spinning larvae, extracted and separated by SDS-PAGE ( 10%)

and the gel was Western blotted using the antiserum for

^35k

protease, CI-8 or SP2. The samples were run in duplicate.

35k protease and for CI-8 at the same position (Fig. 3-20B). The major band was estimated to have a molecular weight of 68,000 to 69,000. The author proposes that this is a 35k protease/CI-S complex. The antiserum for SP2 revealed no strong band at the position. On the other hand, the extract from the unblackened region of midgut fMG-(a+m), non-shaded region in Fig. 3- 20A] had abundant amount of antigen for CI-8 (patterns not shown).

Trial to form a complex of 35k protease and labeled CI -8S in vitro

Purified 35k protease (25 units in I 0 mM Tris-HCl buffer, pH 9) and digestive juice (5]ll) from larvae on day 5 of the 5th instar buffered to pH 9.0 by 0.01 M Tris-HCl were each incubated with 35S-Iabe1ed CI-8S (100 units,

1,000 cprn/rnl) at 30°C for 10 min and the mixtures were subjected to SDS­

PAGE. When the gels were autoradiographed using the Bas 1000 system, a band appeared at a molecular weight of 68,000 to 69,000 (Fig. 3-21). This value was consistent with that obtained in the MG-p extract (cf. Fig. 3-20), indicating that the present enzyme and inhibitor will form a complex not only in vivo but also in vitro.

Localization of CI -8 on the midgut sections using antiserum

On day 5 of the 5th instar, the larval midgut epithelium consists of the three distinct cell types, columnar, goblet and regenerative. The basal side of the entire tissue is covered by a thin basement membrane connecting other tissues like muscles and tracheae which are often intermingled. On cryostat sections, immunological reaction for CI -8 was seen near the basement

membranes of the midgut (Fig. 3-22). Whether CI-8 was present in the cytoplasm of epithelial cells and muscles were uncertain.

126

complex __...

(68-69k Da)

1 2 3

Fig. 3-21. Analysis of a complex of CI-8S and 35k protease formed in vitro.

Digestive juice from larvae on day 5 of the 5th instar (5 ;d, buffered to pH 9.0 by 0.01 M Tris-HCI) and the purified preparation of 35k protease (25 units in 0.01 mM Tris-HCI buffer, pH 9) were each incubated at 30°C for 10 min with 35S-labeled CI-8S (100 units, 1,000 cpm/pl in 0.01 mM Tris-HCI buffer, pH 9) and the mixtures were subjected to SDS-PAGE

(7.4%)

followed by autoradiography using the Bas 1000 system. Lanes 1 and 2, incubated digestive juice and

purified 35k protease, respectively. Lane 3 received untreated 35S-CI-8 S as a marker.

(A)

(B)

Lu

(�) �j:.;�.:.;.:.;.�:.;.��- �;.:.�c c---�1

(B)---��

Lu

� f:Ef:1::t�Itil:::�f�1��j{f :::::�

... •""'-

'1\JIT.J'.J'.l'.l'.l'.J'.J'}')'}' '77

�;::,

Fig. 3-22. Immunohistochemical localization of CI -8 in the midgut from a larva on day 5 of the 5th instar.

Left: Immunostained using anti -CI -8 serum (A).

A control section received normal serum (B). Right: the results are schematically drawn. Scale bar, 50 /.1m. Lu stands for the lumen. For other details see the legend to Fig. 3-18.

{A)

{B)

(A) i<XZZZZWXZIWXZI ^....^.. J l.. ·•

(B) iJVVlJiJ'¢XXXXX > a 1 ··· » ;^.

Fig. 3-23. Immunohistochemical localization of CI-8 in the midgut from a larva on day 2 of spinning.

Left: Immunostained using anti-CI-8 serum

(A).

Specific staining is seen in the disintegrating lMG and M.

A

control section received normal serum

(B).

Right: the results are schematically drawn. Scale bar,

Fig. 3-24. Detailed inspection of the lMG and pMG regions from a larva on day 2 of spinning.

Immunostained using anti-CI-8 serum. Scale bar, 200 Jlffi. For other details see the legend to Figs. 3-18 and 3-22.

130

On day

2

of the spinning stage, the old larval cells were squeezed out into the lumen by newly formed pupal cells, and antigen for CI-8 was detected around the disintegrating larval midgut cells but not in the newly formed pupal midgut cell layer (Figs. 3-23 and 3-24 ).

On day 3 of the pupal stage, the midgut cells became very slender (with increased height and decreased width) and antigen for CI-8 was not observed in the epidermal cells (Fig. 3-25).

Changes in amount of CI-8 in the midgut during development

As shown in Fig. 3-26, the quantity of CI-8 when determined by ELISA started to rise rapidly on days

2

to 3 of the 5th instar, peaking on day

4

or 5.

There was another peak on day 2 of the spinning stage. Thereafter the

quantity dropped keenly until pupation when larval midgut cells were cast off into the 1 ^umen.

Detection of receptor for CI -8S on the midgut and other tissues at various stages of development

From larvae at different ages, peritrophic membrane (PM) fractions were prepared from MG-a and MG-(m+p), extracted and sub

j

ected to SDS­

PAGE, and the gels were ligand-blotted against 35S-labeled CI-8S and autoradiographed using the Bas 1000 system (Fig.

3-27).

A band with a

molecular weight of about 63,000 was recognized strongly on day 5 of the 5th instar, weakly at the onset of spinning and very weakly on day 3 of the 5th instar. This band may represent a CI-8S receptor on the midgut (MG-rec8).

The putative CI-8S receptor like this was scarcely detected in the PM fraction from day

2

spinning larvae (patterns not shown).

w (�

�---� �---�

��: : .

�:. �-Fe�·�- .

^Lu

�i >1:. t�� �

"'!!�! ... :.��- - .. . _-.. ^. %%X"Xnlllo. ...... .

·.:-:

""""""

�

... . !1!11 ... •!1" . . . .

.

^{. . .}

. . . ^.

^.

.

^.

.

^.

^. . .

^{. ... .}

.

^.

.

^,llo.. ... ^)I . . . . . . . . ..... :il

�· .

���� -

�-

· Lu

. .... ·:

^..

^. ,..

^x

..

..

: : :><:

FB ><�-�'-·

• •�: !r•••• · ���

:::

_{.. .. .}

��

^... _?ill

·· · ··::

_�X

��

_{.� ..}

... . .. � � . ^.

.

^�

: :: :

^.

: :

^:

.

: : : ^: : ^: :

�--;:;

. . !-""! ...

: : . ,Ill.: . : : : : : : : : : : : : ·-,

r

.

^..

^.

^.

^.

^.

^.

^{111 • • • . • • • • • •}

.

^10-0-��>-6-(.)..0-6-(.):...o.: u.: : FB: : : : : : : : :

:

^:

I #<><X><><X�· '-'-<:

·

·

>-¢ ·

^>-0-· <><· :x;.

·><>·

<><X><>-6-�._,.. t<::..o<:X:·>-0·

-0-<·

^{�- >0·}

·

<-><· M:· ^.�...:;,-� ^{• • • .• • • • • • • •}

:::::: -� �:: ^:.�

. . . . . . . . Y1lllo

.

^.....

. ^.

^{... '"YYW .}

^..

^.

.

^{... ...}

Fig. 3-25. Immunohistochemical localization of CI-8 in the midgut as well as the fat body from a day 3 pupa.

Top: Immunostained using anti-CI-8 serum

(A).

_Specific

staining is seen in the fat body and in the 1 umen.

A

control section received normal serum (B). Bottom: the results are schematically drawn. Scale bar, 200 Jlffi. For other details see the legend to Figs.

3-18 and 3-22.

C1) ::J en en

� C) E

0

�

_r::

500

300

100

e maleMG --0- female MG

o r��--��--�--��--�--��--�

4M 5-1 5-2 5-3 5-4 5-5 s-1 S-2 P-1 P-2 stage

Fig. 3-26. Developmental changes in the midgut (MG) titer of CI -8 as estimated by ELISA.

The results are indicated in ng/ 10 mg tissue.

5-3 5-5 S-1

1 2 1 2 1 2

�63kDa

Fig. 3-27. Detection of CI-8S receptor in the midgut by ligand assay using 35S-labeled CI-8S.

The membrane fractions of MG-a and MG-(m+p) were prepared from larvae on days 3 and 5 of the 5th in star and on day 1 of spinning (5-3, 5-5 and S-1, respectively). The procedures for extraction, electrophoresis, blotting with labeled CI-8S and autoradiography were as described for the fat body in Chapter 2 (Fig. 2-18). Lanes 1, MG-a; Lanes 2, MG-(m+p ). A positive band at the molecular weight of 63,000 was seen (arrow).

Different larval tissues, e.g., fat body (eFB and gFB, see Chapter 2) on day 2 of spinning, midgut [MG-(m+p)] on day 5 of the 5th instar, testes on day

1 of spinning and ovaries on the day 3 of pupa gave positive sings for the presence of putative CI-8S receptors (Fig. 3-28), indicating that all these organs can incorporate CI-8S. On the other hand, eFB (on day 2 of spinning)

134

did not show positive signs. The bands of gFB, testes and ovaries were at the same position on the gel, thus their receptors were considered to have the same molecular weight, while that for MG-(m+p) seemed to be slightly smaller than others.

FB eFB gFB

MG TS OV

Fig. 3-28. Detection of CI-8 receptor in various tissues.

Detection of CI-8 receptor in eFB, gFB (see Chapter 2), midgut (MG), testis (TS) and ovaries (OV).

The procedures were as described in Fig. 3-27.

Finally, the presence of Cl in n1econium collected from a newly

emerged moth was assessed by zymogram. Only one CI band was revealed at the same position as for purified CI-8 (Fig. 3-29). Upon immunoblotting, this band reacted to the antiserum for CI-8 (patterns not shown). Thus it was identified to be CI -8.

Discussion

Cl-8

Fig. 3-29. Detection of CI in meconium.

Detection of CI in meconiun1 after native PAGE (10%) followed by zymogram analysis.

Arrow indicates the position of CI-8.

There are many studies about proteases of digestive juice and midgut in B. mori (Horie et al., 1963; Horie and Watanabe, 1982; Seki et al., 1982;

Tsuruyama et al., 1985; Tojo eta/., 1986). To obtain information as to a physiological role of CI-8, an endogenous protease which was inhibited by CI- 8 was purified from digestive juice, and the relationship between CI-8 and protease in the midgut during metamorphosis was investigated.

First, the protease purified in this chapter was characterized. Although digestive juice has many kinds of proteases (thus the recovery of purification was very low, since many active fractions were discarded at each of the

136

chromatographic steps), the present enzyme may be one of the major

proteases. On the basis of apparent molecular weight (35,000) it was named 35k protease. Its characteristics including pi (5.1) were different from those of previously reported B. nwri digestive juice proteases (Sasaki and Suzuki, 1982; Eguchi and Kuriyama, 1985; Tsuruyama et al., 1985; Tojo et al.,

1986). Thus 35k protease is a novel molecule. Moreover, the presence of sugar (thus this enzyme is probably anN -glycoprotein) may be the first indication for B. mori digestive juice proteases.

The present protease was not stable at high temperatures but most active under high alkaline conditions at pHs 11 to 12, indicating that it is acting in digestive juice. The activity was inhibited strongly by DFP and PMSF, similarly to those of previously reported digestive proteases (Sasaki and Suzuki, 1982; Seki et al., 1982; Eguchi et al., 1986b), and it seems to be a serine protease. As to the effects of various ions on activity, 35k protease was different from the previously reported three B. nwri digestive juice proteases

(Iwamoto and Eguchi, 1978) whose activities were reduced by Cu2+ (whereas the present protease was unaffected by this ion). On the other hand, three alkaline proteases from the larval gut of an army worm, Spodoptera litura (Ahmad et al., 1980), were not affected by Mg2+, Ca2+, Mn2+ and Co2+ but inhibited by Cu2+, Zn2+ and Hg2+.

Upon SDS-PAGE, crude digestive juice showed dual bands close to 35k protease, both with enzymatic activity and immunological cross-reactivity (thus tentatively called isozymes A and B). The larger one (the B type) showed a molecular weight of about 38,000 to 40,000 (slightly larger than the A type of 35,000) as estimated by SDS-PAGE and it could be named 40k protease

(results are schematically illustrated in Fig. 3-12; details for calibration of B are not shown). The protease activity was stable after incubation of digestive juice at room temperature so that the dual presence seems not to due to an

artifact formed during preparation of digestive juice. Moreover, there were individuals of the cz strain having only 35k protease (the AA type). Thus there is the possibility that these are allelic counterparts.

Among the specimens of the midgut divided according to the region, antigens for 35k protease were detected mainly in the middle region (MG-m) in the case of feeding larva. It is highly plausible that this enzyme is

synthesized in MG-m. Although sharing the same optimum pH (details not shown), the midgut 35k protease may be a precursor which are to be secreted into digestive juice, since the catabolic processes in the midgut cells are

ascribable to peptidases (Horie eta/., 1963, 1982). The anterior region of the midgut (MG-a) seemed to have other digestive pro teases. As to digestive amylases of B. mori, MG-a was the major synthetic site (Kanekatsu, 1980) and some digestive enzymes were also produced here (Kanekatsu, 1980; Santos e _t al., 1984; Azuma, 1995). Large differences in function must exist according

to the location of synthesis. Light and electron microscopical observations (Cioffi, 1979) have indicated that the midgut of M. sexta can also be divided into the anterior, middle and posterior regions on the basis of the folding pattern of the epithelial cell layer and the variation in structure of the goblet and columnar cells.

The present immunoblotting analysis sho\�ed that the putative precursor of 35k protease is embedded in the peri trophic men1branes from which it is released into the lumen. The stepwise degradation and activation of digestive protease were proposed (Eguchi and Iwamoto, 1975, 1976; Iwamoto and Eguchi, 1975, 1978, 1979; Eguchi et al., 1979b, 1986a; Eguchi and

Kuriyama, 1985; Suzuki et al., 1991). Judging from previous observations by microscopy (Cioffi, 1979; Kanekatsu, 1980), the process of protease secretion may not involve normal exocytosis but budding like that seen in dividing yeast.

Newly synthesized protease might be transferred to the apicaltnembrane and

encircled in a large vesicle, which maight be then fused with the peritrophic membrane. Finally protease could be released into the lumen.

The mode of developmental changes in protease activity was similar between the midgut and digestive juice during the early 5th instar. In digestive juice, the proteolytic activity began to drop steeply at the onset of spinning, when silkworm stops eating. At this stage, there were already no proteases in digestive juice, indicating that almost all digestive proteases including 35k protease are discharged into the posterior midgut.

CI -8 was detected near the basement membranes of the midgut epithelial cells adjacent to hemolymph and was not observed in the cytoplasm and the muscle attached the epithelial cells in 5th instar larvae. On the other hand, at the onset of spinning, CI -8 was clearly detected in the cytoplasm of the

degrading larval midgut epithelial cells, while it was not detected in the newly formed pupal midgut epithelial cells.

It is worth noting here that CI-8 strongly inhibited 35k protease. This made the author to examine the problem as to whether CI -8 can meet 35k protease in vivo. After the onset of spinning, the larval midgut undergoes histolysis (accompanying reformation occurring in parallel), and the

proteolytic activity was located in the degraded larval midgut cells but not in the newly formed cells. Proteolytic activity in the degraded cells might participate in the autodigestion of cells. The apparent proteolytic activity decreased as the larval-pupal metamorphosis progressed. Digestive juice proteases did not undergo rapid self-degradation since the original activity was kept after incubation in vitro at room temperature (26.8°C) for 24 hr.

Supposedly, digestive juice pro teases attack the tnidgut cells according to a developmental program leading to apoptosis, which Cis might be involved in.

In fact, CI-8 was also found to enter the midgut tissues from hemolymph before the onset of spinning. It may reach the cytoplasm to control the action

complex with 35k protease in an in vitro experiment (the molecular weight of the putative complex was 68,000 to 69,000, roughly in agreement with the expected value of 77 ,000). A similar complex-like material was detected in the extracts of midgut specimens. The author proposes that CI -8 functions as a modulator of proteolytic activity in the larval midgut tissues during

metamorphosis by making a complex. CI-8 may be incorporated into the larval midgut cells via a CI-8 receptor which was estimated be a 63,000 protein locating on the membrane of the midgut cell.

In summary, it is concluded that CI-8 is accumulated in the larval midgut cells which are to be degraded during metamorphosis and might control the apoptosis by forming a cotnplex with 35k protease.

General Discussion

B. mori has many kinds of hemolymph proteins such as SPs, lipoprotein, vitellogenin and bactericidal peptide etc. Moreover, there are proteinaceous inhibitors active against various proteases (Nakamura, 1981; Yamashita and Eguchi, 1987; Eguchi and Yamamoto, 1988; Yoshida eta/., 1990; Eguchi e t al., 1994). Cis exist abundantly in hemolymph with a lot of variants, and these have been characterized from the viewpoints of genetics (Fujii eta/., 1989; Aratake, 1990; Fujii eta!., 1996a and b), biochemistry (Eguchi and

Kanbe, 1982; Eguchi and Shomoto, 1984; Eguchi ^et al., 1986b; Fujii et al., 1989; Deng et al., 1990) and molecular biology (Sasaki, 1978, 1984; Eguchi and Shomoto, 1985; Sasaki et al., 1990; Matsui and Eguchi, 1991; Narumi et a/., 1993). On the other hand, the present study aimed at approaching the physiological aspects of Cis.

Previous genetical studies cited above (also mentioned in the

introduction for Chapter 1) have confirmed that Cis are controlled by the genes named let-A, Iet-B, /et-D, Jet-E and let-H. Cis in hemolymph of the silkworm are able to be classified into at least three groups: (1) The let-A, Iet-B and /et-H group (gene locus 2-23.7), (2) the Ict-D group (gene locus 19-29.1) and (3) the Jet-E group (gene locus 22-0.0). In the present study, Cls-3, 8 and 13', controlled by different genes, were chosen as representatives of each group.

The characteristics of these Cis examined after purification are summarized as follows: CI-13' had a low molecular weight and acidic pi, similarly to CI -13, an allelic counterpart of CI -13 '. Cls-3 and 8 had higher molecular weights than CI -13'. CI -8 contains an oligosaccharide moiety but

Cis-3 and 13' did not. CI-3 (as well as CI-4 analyzed in hemolymph specimens) had quite different antigenisity from that of CI-8. Cis-6 and 7 (analyzed in hemolymph specimens), which are counterparts of CI-8, reacted with the antiserum for CI-8. Cis-3 and 8 inhibited the proteolytic activity of digestive juice but CI -13' did not. Thus the properties differ from group to group and sometimes from CI to CI within a group, suggesting that Cis have complicated functions.

The developmental changes in concentration of CI-8 antigen was very similar to those of the total inhibitory activity of Cis. Both increased

gradually during the early 5th instar, rapidly during the mid-5th instar, reached maximum at the onset of spinning and then keenly decreased. At maximum, the amount of CI-8 was about 1 mg (2.4 x 10^-8 moles)/ml (about five-fold compared to young larvae). This large extent of changes in titer was taken to imply that CI-8 participate in orne develop1nentally programmed processes rather than self-defence frorn the attack of invading fungi and bacteria. In this connection of interest was the finding that CI-8 decreased steeply from hemolymph after the onset of spinning. CI-8 was sequestered into the fat body like SPs to be stored in granules, and also incorporated into the midgut, testes and ovaries probably to act as a modulator (see below for midgut data). The amounts of CI-8 in these tissues notably changed during the larval-pupal metamorphosis (detailed results for testes and ovaries are not shown).

In the present article, two kinds of fat body tissues were discriminated

according to the location, eFB near the epidennis and gFB around the midgut.

The eFB cells synthesized CI-8 and secreted it into hemolymph during the feeding stage. On the other hand, gFB synthesized little or no CI -8 but

sequestered it from hemolymph into the cytoplasm after the onset of spinning.

In the case of SPs (Tojo et al.,I980; 1981), it has been believed that the same,

multifunctional fat body cells elicit the dual roles of synthesis and sequestration before and after the so-called metaplasia. In this study, eFB and gFB could be discriminated already at the 5th larval instar. During the pupal stage, eFB completely disappeared while gFB thoroughly persisted (as deduced from the observation that the fat body tissues became uniform and the body cavity was filled with gFB-like tissues). This is in agreement with the situation in H. ^zea (Haunerland ^et al.,

1990;

Wang and Haunerland,

1992;

cf. also the citation in Discussion of Chapter

2),

wherein structurally and functionally different fat body tissues are easily distinguished by a blue storage protein. It is certain that B. mori and H. ^zea have two cell types of fat body in an individual. The

present study for B. mori provided the first indication of the functionally different fat body tissues by biochemical and immunohistochemical analyses using a single molecular (CI-8) as a marker. Also similar analyses of eFB and gFB with antisera for the carotenoid binding protein (CBP) and major protein 5 (MP-5) of B. mori hemolymph have drawn the satne conclusions as that for CI-8 (Fujii and Shirai, unpublished data).

Then comes the question as to how the functional difference of the two types of fat body gives rise to. The author inferred that the plasma membranes of gFB had a receptor for CI-8 after the onset of spinning. This was

confirmed to be the case by the ligand assay perfonned by exploiting a

recombinant CI-8 (named CI-8S). This preparation produced by a baculovirus expression system had an advantage in that it was facilely radio-labeled.

Moreover, CI-8S was very close to CI-8 in general properties and authentic enough to be used for experiments as a representative of CI-8. In the extract of gFB, the CI-8 receptor was detected after the onset of spinning but not at the 5th instar. On the other hand, eFB thoroughly showed no CI-8 receptor.

At the mid-pupal stage, the fat body tissues near the epidermis and around the midgut both contained the CI-8 receptor, supporting the idea that eFB was