Disproportionate regulation of nuclear- and mitochondrial-encoded cytochrome oxidase subunit proteins by functional activity in neurons

Vol. 67, No. 1, pp. 197-210, 1995 Elsevier ScienceLtd Copyright © 1995IBRO Printed in Great Britain.All rights reserved 0306-4522/95 $9.50+ 0.00

Neuroscience

~

Pergamon

0306-4522(95)00043-7

DISPROPORTIONATE REGULATION OF NUCLEAR- AND MITOCHONDRIAL-ENCODED CYTOCHROME OXIDASE SUBUNIT PROTEINS BY FUNCTIONAL ACTIVITY IN NEURONS S. LIU and M. W O N G - R I L E Y * Department of Cellular Biology and Anatomy, Medical College of Wisconsin, Milwaukee, WI 53226, U.S.A. Abstraet--Cytochrome oxidase is the terminal enzyme in the mitochondrial respiratory chain engaged in oxidative metabolism and energy production. In mammals, the holoenzyme is composed of 13 subunits encoded by both nuclear and mitochondrial genomes. The goal of the present study was to compare the effect of afferent impulse blockade on the expression of these two genomes at the subunit protein level. It also aimed to determine the correlation between the level of cytochrome oxidase activity and the relative amount of subunit proteins. Relative enzyme activity was analysed hist0chemically, and relative amounts of subunits IV (nuclear-encoded) and I1/III (mitochondrial-derived) proteins were obtained immunohistochemically by anti-subunit IV and anti-subunit II/III antibodies in the lateral geniculate nucleus and the primary visual cortex of adult monkeys. In the normal visual centers, similar staining patterns were found for all three markers. After three and seven days of tetrodotoxin treatment, levels of enzyme activity and subunit proteins declined disproportionately in the deprived laminae of the visual center. Densitometric analysis indicates that changes in enzyme activity and subunit IV proteins were significantly greater than those of subunit II/1II proteins (P < 0.01). The finding that nuclear and mitochondrial genomes are disproportionately regulated at subunit protein levels by neuronal activity implies that the two genomes operate under different regulatory mechanisms. Changes in subunit IV paralleled most closely those of cytochrome oxidase activity (coefficient of determination r 2 = 0.95). This suggests that nuclear-derived subunit IV protein may play a pivotal role in controlling cytochrome oxidase holoenzyme activity.

Cytochrome oxidase (CO), an inner mitochondrial membrane protein, is a critical enzyme in the electron transport chain. Since oxidative energy metabolism is closely coupled to neuronal functional activity, 19'3~ CO has been widely Used as a functional and energy metabolic marker for neurons. ~8'38'4°'4~ The distribution of enzyme activity in the brain is quite heterogeneous at regional, laminar, cellular and subcellular levels, a'~7"~8,3s The level of CO activity in neurons is also regulated by changes in neuronal activity. 12.13.37,42,43 In mammals, the CO holoenzyme is composed of 13 subunits. The three largest suburtits (I-III) are encoded by mitochondrial genes, and the other 10 subunits (IV-VIII) are encoded by the nucleus and are imported into mitochondria. 29'3° The catalytic core of the enzyme is composed mainly of mitochondrially synthesized subunits. The function of the other cytoplasmically synthesized polypeptides is unclear. However, the occurrence of tissue-specific isoforms of some nuclear-encoded subunits (Via, *To whom correspondence should be addressed. CO, cytochrome oxidase; DAB, 3,3'diaminobenzidine; LGN, lateral geniculate nucleus; phosphate-buffered saline; SDS, sodium dodecyl sulfate; TTX, tetrodotoxin.

Abbreviations:

PBS,

VIIa, and VIII) suggests that they confer a regulatory role by adjusting the enzyme activity to the metabolic demands of various tissues. 2,15,~6'z6 Neurons, unlike other cell types, are unique in having processes that extend far away from cell bodies. Mitochondrial genome in the distal processes can, therefore, be separated from the nucleus by a great distance. Little is known about the regulation of gene products from the two genomes in neurons in response to altered functional activity. We previously reported that changes in CO protein level were related to changes in m D N A and CO subunit m R N A s encoded by either mitochondrial or nuclear genes in neurons. 9 Moreover, our in s i t u hybridization studies showed that m R N A s encoded by these two genomes had similar p~tterns of distribution in the normal monkey brain, 8 but were regulated disproportionately by altered neuronal activity. 9 Mitochondrial-derived subunit I m R N A s were down-regulated more severely than nuclear-encoded subunits IV and VIII. However, it is unclear if this is also true at the CO subunit protein level. The present study sought to investigate whether nuclear and mitochondrial genomes are coordinately expressed at subunit protein levels and if the product of one genome is more adversely affected by altered neuronal activity.

197

S. Liu and M. Wong-Riley

198

W e also wished to k n o w the correlation between C O activity a n d C O subunit protein c o n t e n t a n d to further determine w h e t h e r p o s t - t r a n s c r i p t i o n a l control of subunit proteins is also involved in the regulation o f C O activity. Histochemical a n d i m m u n o h i s t o c h e m i c a l approaches allow us to identify changes in C O activity a n d s u b u n i t proteins, respectively, within discrete regions, laminae a n d cell types o f the brain. Therefore, they are the m e t h o d s of choice for the present study.

EXPERIMENTAL PROCEDURES

Immunoblotting Subunit-specific polyclonal antisera used in this study were generous gifts from Dr B. Kadenbach. Human heart CO subunits (II/III and IV) were used as antigens in rabbits. Specificities of these antibodies were tested by western blot, as reported previouslyJ 4 However, whether these antisera could specifically cross-react with monkey brain CO subunits had not yet been examined. In the present study, the specific cross-reactivity of the antisera with monkey brain tissue was tested on western blots. Monkey brain tissues were homogenized in 10% sodium dodecyl sulfate (SDS; l # g tissue/4~tl of SDS) with polypropylene pestles (Kimble) in 1.5 ml polypropylene tubes, and the homogenates were solubilized overnight at room temperature. Samples were centrifuged at 45,000 r.p.m, for 30 min, and the supernatant was diluted with 2 × sample buffer containing 0.125M Tris (pH 6.8), 40% glycerol and 0.01% Bromophenol Blue, and boiled for 5min at 90°C. /~Mercaptoethanol was added to a final concentration of 5% 1 h before electrophoresis. Low molecular weight standards (Bio-Rad) were used as molecular markers in this study. Fifty micrograms of total brain protein was loaded in each lane of an SDS-12% polyacrylamide~ M urea gel, and ran at 50 V until the protein was through the stacking gel, then at 100V at room temperature. Separated proteins were transferred to nitrocellulose in an electroblot apparatus in transfer buffer (25 mM Tris, pH 8.3, 192mM glycine and 20% methanol) at constant 100V for l h at room temperature. The nitrocellulose was dried at 65°C for 30 min and blocked in 5% non-fat dry milk overnight at 4°C in 150 mM Tris buffer, pH 7.4, with 0.89% saline and 0.2% Tween 20. Blots were then incubated with subunit-specific antisera at an optimal antibody dilution (anti-CO II/III, 1:2000; antiCO IV, 1:5000) for 3-4 h at room temperature. The same concentration of normal rabbit immunoglobulins and the deletion of primary antibodies were used as controls. After three washes with Tris-buffered saline-Tween 20, 10min each, blots were reacted with secondary antibodies (Bio-Rad, GaR-horseradish peroxidase) at a dilution of I0,000 in TBST. Immunoreactive bands were visualized by an enhanced chemiluminescence system and exposed to X-ray film according to the manufacturer's instructions (Amersham Life Science).

Animal and tissue preparations Two normal (Macaca mulatta or Macacafascicularis) and two tetrodotoxin (TTX)-treated adult monkeys (Macaca mulatta) were used in this study. Experimental animals were injected with TTX (19/ag/10 ~tl saline) into the left eye intravitreally once or twice a week and survived for three or seven days. Animals were killed by cervical dislocation under deep anesthesia (Ketamine, 20mg/kg, i.m.; and sodium pentobarbital, 45 mg/kg, i.p.). Brains were removed and fixed by immersion in 4% paraformaldehyde, 4%

sucrose and 0,1 M sodium phosphate, pH 7.4, for 6 h at 4°C with constant agitation. Tissues were rinsed three times in cold 0.1 M phosphate-buffered saline (PBS) and 4% sucrose and sustained cryoprotection in 10%, 20% and 30% sucrose in PBS. The lateral geniculate nucleus (LGN) and the primary visual cortex (area 17) were cut coronally or tangentially at 20 # m with a freezing microtome. Frozen sections were collected in 0.1 M PBS, pH 7.4, and adjacent sections were further processed for CO histochemistry and CO subunit immunohistochemistry.

Cytochrome oxidase histochemistry and cytochrome oxidase subunit immunohistochemistry Reaction product of CO activity and immunoreaction product of nuclear-encoded CO subunit IV and mitochondrial-encoded subunits II/III will be referred to as the three markers. To obtain more accurate results for these markers, we carried out the following. (1) The optimal dilution for each subunit antiserum was determined by titration on monkey LGN and visual cortex sections. Final antibody dilutions were used at a concentration beyond which specific immunostaining would not be enhanced, but non-specific background staining would increase. (2) The optimal incubation times for both CO histochemistry and immunohistochemistry were also determined by repeated trials, and coincided with the time just before the reaction began to plateau. The optimal antibody dilutions and 3,3'diaminobenzidine (DAB) reaction time for each brain region were found to be reproducible between animals. (3) Since antisera possessed different affinities to their antigens, the level of staining intensity for different subunit antisera would not be the same. For this reason, only the proportional change in staining intensities between deprived and non-deprived LGN and visual cortex of each animal was determined for each marker and compared between the three markers. CO histochemistry was performed as described previously. 37 In brief, sections were incubated in a 0.1 M phosphate-buffered (pH 7.35) reaction medium containing 50mg% DAB and 60mg% cytochrome c (Sigma) for approximately 4 h at 37°C with agitation. Following the reaction, sections were washed three times in 0.1 M PBS, pH 7.35. Free-floating sections were processed for immunohistochemistry. They were blocked overnight at 4°C in 0.1 M phosphate buffer, pH 7.4, containing 5% non-fat dry milk, 5% normal goat serum and 0.5-1% Triton X-100. The sections were placed in one of the following primary subunit-specific antibodies: antiserum to human heart CO subunits II/III diluted 1:3000 or antiserum to subunit IV diluted 1:10,000 in the blocking solution. Sections were reacted with primary antibodies for 4 h at room temperature, then for 24 h at 4°C. They were then incubated with secondary antibodies (Bio-Rad blotting grade GaR-horseradish peroxidase) at hl00 dilution in the same solution overnight at 4°C. Immunoreaction product was visualized by 0.05 mg% DAB and 0.004% HzO 2 for 5-10 min at room temperature. In the control experiment, sections were incubated with either normal rabbit immunoglobulin G diluted at the same concentration as primary antibodies or with no primary antibodies and were processed as above. Immunostained sections were mounted on gelatin-subbed slides, air dried, dehydrated in ethanol, cleared in xylene and coverslipped with Perrnount.

Quantitative densitometry of normal and tetrodotoxin-treated monkey lateral geniculate nucleus Computer-assisted quantitative optical densitometric measurements of reaction product following CO histochemistry or subunit immunohistochemistry were used to assess levels of CO activity and the relative amounts of subunit proteins, respectively, in each lamina of the LGN and layer 4c of the primary visual cortex of normal and TrX-injected

199

Regulation of cytochrome oxidase subunit proteins in neurons monkeys. The Microcomputer Imaging Device system (Imaging Research, ST. Catharines, Ont.) with eight-bit resolution per pixel was used. Images of the LGN and the primary visual cortex in coronal sections were taken by a SIT video camera (MTI model PA-70) attached to a Leitz microscope, and the output was analysed by a Perceptics Model 9200 Image Processor by means of HYPERSCOPE and IPSERVER (Biovision) programs on a Macintosh IIfx computer. For all three mai'kers, measurements were taken from cell bodies and neuropil of all six layers of the most central portions of the left LGN in normal monkeys and in monkeys whose left eyes were injected with TTX. Optical densities of the visual cortex were taken from light and dark columns in cortical layer 4c. Each reading was obtained in a circular region of 10 x 10 pixels, and approximately 75-100 readings from each lamina were taken for each of the three markers. Since sections processed for different markers showed varying levels of background staining, each optical density reading, in turn, was divided by an average optical density value of the white matter in the same section. Therefore, the relative optical density (a ratio of optical densities of gray matter versus white matter) was used to represent staining intensity for all three markers and for all of the quantitative analyses and comparisons. Laminar comparisons of relative optical densities for each marker were made only on the same section to ensure that all laminae were processed under the same conditions. For each marker, comparisons were performed between non-deprived (1, 4 and 6) and deprived laminae (2, 3 and 5) of LGN and between normally innervated and deprived ocular dominance columns of layer 4c in normal and TTX-treated monkeys. We also compared changes in staining intensities following TTX treatment between nuclearencoded subunit IV and mitochondrial encoded subunits II/III, and between CO activity and the immunoreactivity of subunits IV and II/III. Statistical analysis was done using two-tailed Student's t-test for the comparison of relative optical densities of each marker between non-deprived and deprived laminae or columns of the LGN and the visual cortex, and one-way analysis of variance for comparisons among different markers for changes in staining intensities. In both cases, a P value of less than 0.05 was considered significant. RESULTS

Specificity of antisera to cytochrome oxidase subunits in the monkey brain Antisera used in this study were generated against human heart C O subunits II/III or IV and were characterized by western blots, ~4 which showed that they reacted only with the corresponding subunits of human heart CO. However, antibodies are not always feasible for use in species different from the one in which antigens are obtained. Our western blots showed that both subunit-specific antisera crossreacted only with their corresponding subunit proteins in monkey brain tissue (Fig. 1, lanes 1 and 2). N o r m a l rabbit immunoglobulins gave background staining (Fig. 1, lane 3). The antibodies were therefore considered valid for our immunohistochemical studies of the monkey brain.

Distribution of cytochrome oxidase activity and cytochrome oxidase subunit immunoreaetivity in normal monkey lateral genieulate nucleus The normal monkey L G N is composed of six cellular layers separated by fibrous bundles. Laminae

1

2

3

Mr xlO"3.

45-

3121.5-

Fig. 1. Homogenates of monkey brain tissues were run on a 12% SDS-polyacrylamide~M urea gel. Fractionated proteins were transferred to nitrocellulose membranes that were subsequently incubated with polyclonal antisera either to subunits II/III (lane 1, Mr: 26,500) or subunit IV (lane 2, Mr: 17,500). Antisera were diluted 1:2000 (lane 1) and 1:5000 (lane 2). Immunoreactive bands were visualized on western blots by the enhanced chemiluminescence method. Both sets of polyclonal antibodies showed monospecific cross-reactivity with their corresponding polypeptides in the monkey brain. Normal rabbit immunoglobulins (diluted at 1:2000) were used as a control (lane 3). 2, 3 and 5 receive retinal input from the ipsilateral eye, while laminae 1, 4 and 6 receive input from the contralateral eye. The laminar distribution and cellular localization of CO activity in the normal monkey L G N were in agreement with our previous findings. 2j,3s Briefly, intense staining was present in all six laminae, but magnocellular laminae 1 and 2, and often parvicellular lamina 6, were more darkly stained for CO histochemistry than the other parvicellular laminae (3, 4 and 5; Fig. 2A). CO-reactive and non-reactive neurons were observed throughout six layers of the L G N (Fig. 2A). Laminar and cellular immunostaining patterns for nuclear-encoded subunit IV and for mitochondrial-encoded subunits I1/III appeared identical and were similar to the pattern of CO activity (Fig. 2A-C). Immunoreaction product was present in both cell bodies and neuropil throughout the L G N (Fig. 2B, C). CO histochemistry gave more intense staining, a lower background and

200

S. Liu and M. Wong-Riley

Fig. 2. Coronal sections through the LGN of a normal adult monkey showing distribution patterns of CO activity and CO subunit proteins. Sections were processed for CO histochemistry (A), immunohistochemistry of subunit IV (B) and subunits II/III (C). There is clear immunoreactivity of the monkey LGN with antisera against human heart subunits IV and II/III, and a similar staining pattern is found for all these markers. Intense CO activity and immunoreactivity can be seen in all six laminae, especially laminae 1 and 2. The same concentration of normal rabbit immunoglobulin G (D) gives a low level of non-specific background. Scale bar = 1 mm.

a higher contrast than subunit immunohisochemistry (Fig. 2A-C). On the other hand, subunit immunostaining more prominently labeled neuronal cell bodies than the neuropil (Fig. 2A-C). The same concentration of normal rabbit immunoglobulin G exhibited only faint background staining (Fig. 2D).

Tetrodotoxin-treated nucleus

monkey

lateral

geniculate

In contrast to normal monkey LGN, after three days of monocular TTX injection, staining intensity was reduced for all three markers in the deprived laminae 2, 3 and 5 ipsilateral to the TTX injected eye (not shown). The reduced staining was found in both cell bodies as well as in the neuropil of magnocellular and parvicellular layers. No significant cell loss was detected in deprived laminae. Reductions in staining intensities for all three markers were more severe in deprived laminae after seven days of TTX treatment, as indicated by asterisks in Fig. 3. Nuclear-encoded

subunit IV showed a greater reduction in deprived laminae than mitochondrial-derived subunits II/III at both time periods, though the reduction was more severe after seven days of TTX. Moreover, the change in subunit IV immunoreactivity paralleled closely that of CO activity (Fig. 3). Deprived ipsilateral laminae were affected more by TTX treatment than deprived contralateral laminae (not shown).

Quantitative analysis In the normal monkey LGN, quantitative analysis indicated that the level of CO activity (Fig. 4, time 0, left panel) and relative amounts of subunit IV proteins (Fig. 4, time 0, middle panel) and II/III proteins (Fig. 4, time 0, right panel) were slightly greater in layers 1, 4 and 6 (grouped together and normalized in each panel to yield an optical density value of 1; solid circles) than in layers 2, 3 and 5 (grouped together and expressed as relative optical densities, open circles, with reference to the contralateral

Regulation of cytochrome oxidase subunit proteins in neurons laminae). For each marker, the discrepancies in staining intensity between contralateral and ipsilateral laminae were statistically significant (P < 0.05). This is consistent with our previous findings,2~ indicating that the levels of metabolic activity are higher in contralateral laminae (1, 4 and 6) than ipsilateral ones (2, 3 and 5) in the normal monkey LGN. In addition,

201

there was no significant difference in staining intensities among all three markers (P > 0.05). Three days of TTX treatment brought about a greater difference in staining intensities between contralateral (non-deprived) and ipsilateral (deprived) laminae. Such differences exceeded those of the normal L G N by 3.l-fold for CO activity (Fig. 4, time

Fig. 3. Distribution of CO activity and amount of CO subunit proteins (IV and II/III) in left LGN of a monkey whose left eye was inactivated by TTX for seven days. (A) CO activity; (B) subunit IV; and (C) subunit lI/III immunoreactivity. Staining intensities declined in deprived laminae 2, 3 and 5 (asterisks) ipsilateral to the experimental eye for all three markers as compared to normal ones (see Fig. 2). Changes in CO activity and subunit IV immunoreactivity are clearly more severe than those of subunits II/llI. Scale bar = I ram.

S. Liu and M. Wong-Riley

202

c-

CO II/111

CO IV

CO activity

'ot'i

1.o

"0 0.8 o 00

._> n-

0.6

0.6

0.4

. 0

.

.

. 3

.

.

i 7

0.4

0.6

,

,

~

, 3

1

~

1

0 7

. 0

4

=

~

,

~

1

3

7

TTX Treatment Time (days)

Fig. 4. Quantitative comparisons of differences in staining intensities between non-deprived and deprived LGN laminae for all three markers (CO activity, subunit IV and subunit II/III immunoreactivity). Optical densitometric measurements were made by means of a computer-assisted image processor, and the means ( _ S,E.) of 75-100 readings from each LGN lamina, including cell bodies and the neuropil, were represented at each time point for each marker. Each of these readings was divided by a reference value (white matter) in the same section so as to yield a relative optical density value. Data were then grouped such that the non-deprived laminae (in this case, laminae 1, 4 and 6; solid circles) were normalized to yield a density of 1, and deprived laminae (in this case, laminae 2, 3 and 5; open circles) showed changes relative to the non-deprived laminae. The graphs show that there are significant differences in staining intensities between contralateral (1, 4 and 6) and ipsilateral (2, 3 and 5) laminae in the normal monkey LGN for all three markers represented in the figure (small asterisk, *P < 0.05 at time 0). Decreases in deprived ipsilateral laminae are about 18% for CO activity (at time three days; left pane|), 17% for subunit IV (middle) and 14% for subunits II/III (right) after three days of TI'X (large asterisk, *P < 0.001), and 23 %, 25% and 18%, respectively, after seven days of TTX (at time seven days; large asterisk, *P < 0.001). CO activity and subunit IV are affected 1.3- to 1.4-fold more than subunits II/III by three days, and 1.4- to 1.5-fold by seven days of TTX. Significant differences exist between CO activity and subunits II/III and between subunit IV and subunits II/III after three and seven days of TTX (P < 0.05 and 0.01, respectively).

three days, left panel), 2.8-fold for subunit IV (middle panel) and 2.l-fold for subunits II/III (right panel; P <0.001 for all). Changes in CO activity and nuclear-derived subunit IV proteins were significantly greater than those in mitochondrial-derived subunit II/III proteins (P < 0.05). There was no significant difference between changes in CO activity and subunit IV (P > 0.05). Thus, quantitative analysis coincided with our qualitative observations, indicating that CO activity and subunit IV proteins were affected more severely than subunit II/III proteins after three days of TTX. After seven days of T T X inactivation, differences in staining intensities between contralateral (nondeprived) and ipsilateral (deprived) laminae exceeded those of normal L G N by 4.1-fold for CO activity (Fig. 4, time seven days, left panel), 4-fold for subunit IV (middle panel) and 2.7-fold for subunits II/III (right panel; P < 0.001 for all). Changes in staining intensities were greater for CO activity (by 1.5-fold) and for subunit IV proteins (by 1.4-fold) than for mitochondrial-encoded subunit II/III proteins

(P < 0.01 for both). No significant difference was found between CO activity and subunit IV proteins (P > 0.05).

Correlation between cytochrome oxidase activity and subunit protein levels Relative optical densities of immunohistochemical staining for subunits IV and II/III were plotted against those of CO histochemical staining from adjacent sections of the left LGN. Linear regression analysis revealed that, in the normal LGN, there was a positive correlation between CO activity and subunit IV (r ~ = 0.924; Fig. 5A) and between CO activity and subunits II/III (r 2--- 0.942; Fig. 5B). However, following T T X treatment, the decline in the a m o u n t of subunit IV correlated highly with the decrease in CO activity (r 2 = 0.950 for three days and r 2 = 0.915 for seven days of TTX; Fig. 5C and E, respectively). The coefficients of determination between subunits II/III and CO activity were only 0.774 and 0.721 for three and seven days of T T X (Fig. 5D and F, respectively).

Regulation of cytochrome oxidase subunit proteins in neurons

NORMAL

A. 0.30

B.

NORMAL

0.30

2

O

r2=

0.26

._>

13 t~

13

O o

O O

0.26

0.22

0.22

.

0.18 ' " 0.20

,

I

.

•

•

•

0.25

!

•

•

0.18

•

,

,

,

C°

13

,

0.20

0.30

I

=

,

,

•

0.25

I

.

0.30

CO II/111protein

CO IV protein

.__>,

203

3 DAYS TTX

3 DAYS TTX

D.

1.3

1.3

1.2

1.2

1.1

.g, 13

1.0

O

1.1 1.0

O 0.9

0.9

0.8

0.8

0.7 / 0.6

,

I

,

I

0.8

~

1.0

I

0.7 0.4

1.2

O

I

I

I

0.5

0.6

0.7

CO IV protein E°

0.8

CO II/111protein

F.

7 DAYS TTX

7 DAYS TTX

1.4 1.0 1.3

:>, .>

13

1.2

0 (..)

O O 1.1

1.0 1.0

0.9

0.8

0.7

I

I

I

1.1

1.2

1.3

CO IV protein

.4

0.6 1.0

I

1.1

O

|

1.2

•

I

1.3

,

1.4

CO II/111protein

Fig. 5. Correlation between CO activity and subunit proteins (IV and II/III) in LGN of normal and TTX-treated monkeys. Each individual point is an average value which represents the difference in relative optical densities between contralaterally and ipsilaterally innervated LGN laminae of normal (A, B) and TTX-injected (C-F) monkeys. Quantitative analyses show that, in normal monkey LGN, the differences between contralateral and ipsilateral laminae are present for all three markers, and a close correlation is found between CO activity and subunit IV (A) and between CO activity and subunits II/III (B). After three or seven days of TTX, greater correlation exists between CO activity and subunit IV (C, E) than between CO activity and subunits II/III (D, F).

204

S. Liu and M. Wong-Riley

Normal visual cortex. In the normal striate cortex, a distinct pattern of CO histochemical staining was observed (Figs 6A, 7A), as described in our previous studiesY 8 A similar distribution was also seen immunohistochemically for CO subunits IV and II/III in sections cut coronally (Fig. 6) or tangentially (Fig. 7). CO-rich supragranular "puffs" or patches 3'12'13were darkly stained histochemically and immunohistochemically. These "puff" regions have specialized physiological properties23'33 and a higher level of energy metabolism 12'13'39'42'43 than the adjacent interpuff regions (Figs 6, 7). Layers 4a and 4c receive afferent input mainly from the L G N and have the highest levels of CO activity and subunit immunoreactivity. Layer 4a had a honeycomb-like pattern and layer 4c showed intense staining for all three markers (Figs 6, 7).

Layer 6 was moderately reactive, while layer 5 was only lightly stained (Fig. 6). Visual cortex o f tetrodotoxin-treated monkeys Three days of unilateral intravitreal injections of TTX caused decreases in CO activity and immunoreactivity of subunits IV and .II/III proteins in alternate rows of puffs representing the manipulated eye (not shown). Ocular dominance columns in layer 4c showed alternating dark and light staining bands, corresponding to normal and TTX-treated eyes, respectively. In general, changes in layer 4c occurred earlier and were more prominent than those in puffs. Some mild effects in CO activity and subunit proteins were also observed in layers 5 and 6 as bands of lighter staining. Greater changes in visual cortex were seen after seven days of TTX inactivation (Fig. 8) and

A

B

C

~ i( i :¸ Fig. 6. CO histochemistry (A) and immunohistochemistry for subunits IV (B) and II/III (C) exhibit similar staining patterns in the normal monkey visual cortex. Puffs (arrowheads) in layers 2 and 3 have higher levels of CO activity and subunit immunoreactivity than adjacent "interputV' regions. Layers 4a and 4c, which receive afferent input mainly from the LGN, also show intense staining by these markers. Layer 6 is moderately stained, while layer 5 is only lightly staine& Scale bar = 0.5 ram.

Regulation of cytochrome oxidase subunit proteins in neurons

205

Fig. 7. The distribution of CO activity (A) and immunoreactivity for CO subunit IV (B) and II/III (C) proteins in tangential sections through area 17 of a normal monkey. Note that layer 4c shows uniform staining for CO histochemistry and subunit immunohistochemistry, layer 4a has a honeycomb appearance, and the arrowheads show puffs in layers 2 and 3. Scale bar = 0.5 mm.

the effect was even more dramatic in tangential sections (Fig. 9). At both time periods, nuclear-encoded subunit IV proteins were reduced to a similar degree as CO activity and both were affected more than mitochondrial-derived subunit II/III proteins (Figs 8, 9). Quantitative analyses of the three markers were also performed in layer 4c of visual cortex from normal and TTX-treated monkeys. There were no differences in staining intensities between ocular dominance columns representing each eye (P > 0.05) in normal monkeys for all three markers. Staining intensities were markedly reduced in deprived columns by 13% for CO activity (P < 0.002), 15% for subunit IV (P < 0.001) and 9% for subunits II/III (P < 0.02) after three days of TTX treatment, and 21%, 22% and 13% respectively after seven days of TTX (P < 0.001 for all). Changes in CO activity and subunit IV protein were 1.5- and 1.7-fold greater, respectively, than those of subunits II/III (P < 0.01 for both) after seven days of TTX injection. There was no significant difference between CO activity and CO subunit IV after three or seven days o f TTX treatment. DISCUSSION

Our western blots showed that the antisera specifically cross-reacted with their corresponding subunit proteins in the monkey brain, and immunohistochemical patterns of both subunit antisera were similar to those of CO activity. The same concentration of normal rabbit immunoglobulins only exhibitedbackground staining. This result confirmed the validity of applying antibodies against human heart CO subunits II/III or IV to the monkey brain. This is not NSC 67/1--H

surprising because the three subunits are highly conserved among species and organs, and no isoforms have yet been reported for them. 2°,35 CO subunit immunohistochemistry has been successfully used in determining the relative amounts of CO subunit proteins in skeletal muscles of patients with partial CO deficiencies) 4.25 The present study further confirms that CO subunit immunohistochemistry could serve as a useful method for studying the expression of nuclear and mitochondrial genes at the subunit protein level in brain tissue.

Distribution of subunit IV (nuclear) and H/Ill (mitochondrial) proteins in normal monkey lateral geniculate nucleus and primary visual cortex It has been reported that, at steady states, nuclear and mitochondrial genomes are proportionately regulated for C O subunit mRNAs. This is evident from studies where the ratio of CO III mRNA to CO VIc mRNA is equal in each rat tissue analysed by cDNA probes hybridized to corresponding mRNAs on slot blots. 7:° Moreover, our previous studies using in situ hybridization have found that mRNAs of subunits I, IV and VIII have a similar pattern of distribution in the monkey LGN and the primary v i s u a l cortex. 9 The present study demonstrates that staining patterns in these visual centers were similar for all three markers (CO activity, CO subunit IV and II/III proteins). These findings indicate that levels of CO activity in varying brain regions, laminae and cell types are normally controlled by amounts of CO subunit proteins encoded by both genomes. However, it remains uncertain whether similar patterns of distribution necessarily imply that the two genomes are coordinately expressed at the subunit protein level.

206

S. Liu and M. Wong-Riley

A,

i I i? ii ~ilill¸

B

C

tV ~

It/Ill

Fig, 8. After seven days of TTX treatment, the normally innervated puffs in layers 2 and 3 and ocular dominance columns in layer 4c are intensely stained, while the deprived ones (open arrows) are lightly stained. CO activity and subunit IV immunoreactivity decreased more than subunits II/III in deprived columns. Scale bar = 0.5 mm.

Nuclear and rnitochondrial genomes are disproportionately regulated at the eytochrome oxidase subunit protein level by altered neuronal activity One of the major findings in this study is that CO subunit IV protein declined by 1.4- and 1.7-fold greater than subunit II/III proteins (P < 0.01) in the LGN and the primary visual cortex, respectively, following seven days of TTX. This indicates that proportionate expressions of CO genes in the steady state were disrupted by altered functional activity, and that nuclear and mitochondrial genomes are disproportionately regulated at the subunit protein level. The disproportionate regulation of the two genomes is also found to be true in skeletal muscles

of patients with partial CO deficiencies, but in that study mitochondrial-encoded subunit II protein was affected more markedly than nuclear-encoded subunit IV. z5 The inconsistent results between neurons and muscles can be explained by the fact that mitochondrial DNA deletions in muscles caused a depletion of mitochondrial-encoded CO subunits, and that was the primary event for CO deficiencies in that case. It is also possible that tissue-specific regulatory mechanisms for the expression of the two genomes are different among organs. This is supported by a recent study which showed that, after thyroid hormone (T3) treatment, the responses of the two genomes were entirely different in rat liver and skeletal muscle. 36 In the liver, nuclear-encoded CO

Regulation of cytochrome oxidase subunit proteins in neurons subunit transcripts as well as proteins (IV, Va, and VIc) were more dramatically affected than mitochondrial-encoded ones (II and III), while the opposite result was observed in muscle. The predominant expression of tissue-specific isoforms of thyroid hormone receptors in each tissue could be one of the reasons for strikingly different responses of the two tissues. These studies, together with our present findings, suggest that modulations of transcriptional and/or translational controls may be via tissueand/or cell type-specific factors. CO biosynthesis is a very complex process not only because it is involved in the expression of both nuclear and mitochondrial genomes, but also because transcription factors and other regulatory factors required for gene expression differ between the two genomes. 36 Moreover, regulatory elements in the pr.omotor regions of nuclear-encoded subunit genes are different from each other, 1,32'35'45Therefore, the regulation of nuclear and mitochondrial CO genes in response to changes in energy demands may not be sufficiently explained by universal mechanisms common to all CO subunit genes. CO subunits could be regulated independently from one another, and the same subunit may be regulated differently among tissues or cell types. In addition, subunit transcripts and proteins could also be regulated differently. It does not seem efficient for a cell to regulate expression of all 13 subunit genes to the same degree in order to meet energy demands. To best understand CO biogenesis subsequent to altered cellular functional activity, one needs to look at the expression of all 13 CO

207

subunits at m R N A as well as protein levels. No evidence from the current study suggests that one genome is selectively affected earlier than the other because both nuclear- and mitochondrial-encoded subunit proteins declined after three days of TTX. This possibility still remains to be determined by time course studies. Little is known about the molecular basis for the regulation of the two genomes that adjusts the level of CO activity to meet altered energy demands in neurons. Our recent in situ hybridization study showed that subunit I m R N A was affected more than subunit IV and VIII mRNAs, and subunit IV m R N A was affected the least among the three messages following TTX treatment in the monkey L G N and the primary visual cortex. 9 This suggests that pretranslational mechanisms apparently operate on the expression of CO holoenzymes in neurons? The present immunohistochemical study showed that CO subunit IV proteins decreased more dramatically than mitochondrial-derived subunit II/III proteins in the monkey L G N and visual cortex after TTX treatment. The difference in the adjustment of m R N A and protein levels of CO subunits implies that subunit protein synthesis and/or degradation may also play an important role in controlling th e expression of CO enzyme in neurons. This finding is also consistent with previous observations in muscles. Chronic low-frequency stimulation of rat muscle induces a 2-fold increase in mRNAs and a 3--4-fold change in enzyme activity, suggesting that mechanisms operating at levels beyond transcription were additionally

Fig. 9. The distribution of CO activity (A) and immunoreactivity for subunits IV (B) and II/III (C) in serial tangential sections through area 17 of a TTX-treated monkey. After seven days of TI'X, staining in layer 4c is divided into alternating light (deprived; large open arrows) and dark (non-deprived) stripes for all three markers. Small open arrows point to rows of pale puffs representing the TTX-injected eye, and arrowheads show adjacent rows of normal puffs. Again, changes in staining intensity for CO activity and subunit IV are more profound than those for subunits II/III. Scale bar = 0.5 mm.

208

S. Liu and M. Wong-Riley

responsible for increased CO activity, t~ Thus, neurons do not simply alter the rate of CO subunit m R N A synthesis, but also regulate protein translation and/or degradation for some or all subunit proteins in order to significantly affect CO activity. Cytochrome oxidase subunit I V could be an important protein for the regulation o f cytochrome oxidase activity

The most important feature provided by our data (see Fig. 6) is that changes in CO subunit IV show the greatest degree of correlation with decreased enzyme activity following TTX treatment. This suggests that subunit IV may play a regulatory role in controlling CO holoenzyme activity in neurons. Recent studies with yeast CO have revealed that deletion mutations in the structural genes for subunit IV, Va plus Vb, and VI led to a complete loss of CO activity, 5"27'34'~ indicating that these subunits are essential for CO holoenzyme function. A study by Merle et al. 24 revealed that CO subunit IV was tightly associated with 1 1 other CO subunit components, which could not be removed by either Triton X-100 (during immunoprecipitation) or by cholate (during isolation) in rat liver. This suggests that subunit IV may be important in assembling CO holoenzymes in mammals, as in yeast. At physiological concentrations, subunits IV and VIII of the CO holoenzyme provide the binding sites for ATP. 2,26 It has long been known that ATP has a regulatory role on the function of CO holoenzyme in yeast as well as in bovine heart. 6,28 Further studies have confirmed that ATP binds to CO holoenzyme to induce an allosteric conformational change, thus affecting the interaction of the enzyme with cytochrome c. 2'26 It is clear from these previous studies that nuclear-encoded CO subunit IV takes part in regulating CO activity by binding to ATP. We recently found that the precursor proteins of CO subunit IV were overproduced in the rat brain and had a similar distribution pattern as that of its

mature form and as that of CO activity. The precursor pool was down-regulated by monocular enucleation. 22 It is possible that controlling the maturation of subunit IV precursor protein is important for the regulation of CO activity. Taken together, the fact that the decrease in nuclear-derived subunit IV was tightly coupled to the decrease in CO activity after TTX treatment leads to the speculation that the decrease in CO activity may, in part, be caused by a deficiency in subunit IV proteins, which could subsequently affect holoenzyme binding to cytochrome c. The precise role of subunit IV in regulating CO holoenzyme level remains to be identified.

CONCLUSIONS

The observations from the present study permit several conclusions: (i) CO subunit immunohistochemistry is a useful method for studying the regulation of CO gene expression at the subunit protein level; (ii) CO subunit proteins from nuclear and mitochondrial genomes are disproportionately regulated by altered neuronal activity in L G N and primary visual cortex of adult monkeys; (iii) posttranscriptional modulations of CO subunit gene expression play an important role in the regulation of CO activity; (iv) changes in the level of CO subunit IV parallel most closely those of CO activity after afferent impulse blockade; and (v) nuclear-derived CO subunit IV may play a regulatory role in controlling the activity of CO holoenzyme.

Acknowledgements--We

arc

very

grateful

to

Dr

B.

Kadenbach for providing us with subunit-specific antisera to human heart subunits lI/III and IV. We thank Dr R. F. Hevner for helpful discussion and advice, Dr T. Trusk for assisting with the use of the Perceptics image processor, Dr M. A. Mullen for helpful suggestions, and Z. Huang and F. Nie for assistance in the Laboratory. This work was supported by grants (NS18122 and EY05439) from the National Institutes of Health.

REFERENCES

I. Basu A. and Avadhani N. G. (1991) Structural organization of nuclear gene for subunit Vb of mouse mitochondrial cytochrome c oxidase. J. biol. Chem. 266, 15450-15456. 2. Bisson R., Schiavo G. and Montecucco C. (1987) ATP induces conformational changes in mitochondrial cytochrome c oxidase. J. biol. Chem. 262, 5992-5998. 3. Carroll E. W. and Wong-Riley M. T. T. (1984) Quantitative light- and electron-microscopic analysis of cytochrome oxidase-ricb zones in the striate cortex of the squirrel monkey. J. comp. Neurol. 222, I-7. 4. Difiglia M., Graveland G. A. and Schiff L. (1987) Cytochrome oxidase activity in the cat caudate nucleus: light- and electron-microscopic observations. J. comp. Neurol. 255, 137-145. 5. Dowhan W., Bibus C. R. and Schatz G. (1985) The cytoplasmically made subunit IV is necessary for assembly of cytochrome c oxidase in yeast. Eur. molec. Biol. Org. J. 4, 1979-1984. 6. Ferguson-Miller S., Brantigan D. L. and Margoliasb E. (1976) Correlation of the kinetics of electron transfer activity of various eukaryofic cytochrome c with binding to mitochondrial cytochrome c oxidase. J. biol. Chem. 7,51, 1104-I I 15. 7. Gagnon J., Kurowski T. T., Wiesner R. J. and Zak R. (1991) Correlations between a nuclear and a mitochondrial mRNA of cytochrome c oxidase subunits, enzyme activity and total mRNA content, in rat tissues. Molec. cell. Biochem. 107, 21-29. 8. Hevner R. F. and Wong-Riley M. T. T. (1991) Neuronal expression of nuclear and mitocbondrial genes for eytochrome oxidase (CO) subunits analyzed by in situ hybridization: comparison with CO activity and protein. J. Neurosci. II, 1942-1958. 9. Hevner R. F. and Wong-Riley M. T. T. (1993) Mitochondrial and nuclear gene expression for cytochrome oxidase subunits are disproportionately regulated by functional activity in neurons. J. Neurosci. 13, 1805-1819.

Regulation of cytochrome oxidase subunit proteins in neurons

209

I0. Hood D. (1990) Coordinate expression of cytochrome c oxidase subunit III and VIc mRNAs in rat tissue. Biochem. J. 269, 503-506. 11. Hood D. A., Zak R. and Pette D. (1989) Chronic stimulation of rat skeletal muscle induces coordinated increases in mitochondrial and nuclear mRNAs of cytochrome c oxidase subunits. Eur. J. Biochem. 179, 275-280. 12. Horton J. C. (1984) Cytochrome oxidase patches: a new eytoarchitectonic feature of monkey visual cortex. Phil. Trans. R. Soc. Lond. 304, 199-253. 13. Horton J. C. and Hubel D. H. (1981) Regular patchy distribution of cytochrome oxidase staining in primary visual cortex of macaque monkey. Nature 292, 762-764. 14. Johnson M. A., Kadenbach B., Droste M., Old S. L. and Turnbull D. M. (1988) Immunohistochemical studies of cytochrome oxidase subunit in skeletal muscle of patients with partial cytochrome oxidase deficiencies. J. neurol. Sci. 87, 75-90. 15. Kadenbach B. and Merle P. (1981) On the function of multiple subunits of cytochrome c oxidase from higher eukaryotes. Fedn. Eur. biochem. Socs. Lett. 135, 1-I 1. 16. Kadenbach B., Khun-Nentwig L. and Buge U. (1987) Evaluation of a regulatory enzyme: cytochrome c oxidase (complex IV). Curr. Top. Bioenerg. 15, 113-161. 17. Kageyama G. H. and Wong-Riley M. T. T. (1982) Histochemical localization of cytochrome oxidase in the hippocampus: correlation with specific neuronal types and afferent pathways. Neuroscience 7, 2337-2361. 18. Kennedy H., Bullier J. and Dehay C. (1985) Cytochrome oxidase activity in the striate cortex and lateral geniculate nucleus of the new born and adult macaque monkey. Expl Brain Res. 61, 204-209. 19. Krnjevic K. (1975) Coupling of neuronal metabolism and electrical activity. In Brain Work: the Coupling of Function, Metabolism, and Blood Flow in the Brain (eds Ingvar D. H. and Lassen N. A.), pp. 65-78. Academic Press, New York. 20. Kuhn-Nentwig L. and Kadenbach B. (1985) Isolation and properties of cytochrome c oxidase from rat liver and qualification of immunological differences between isoenzymes from various rat tissues with subunit-specific antisera. Eur. J. Biochem. 149, 147-158. 21. Liu S. and Wong-Riley M. (1990) Quantitative light- and electron-microscopic analysis of cytochrome-oxidase distribution in neurons of the lateral geniculate nucleus of the adult monkey. Vis. Neurosci. 4, 269-287. 22. Liu S. and Wong-Riley M. (1994) Nuclear-encoded mitochondrial precursor protein: intramitochondrial delivery of dendrites and axon terminals of neurons and regulation by neuronal activity. J. Neurosci. 14, 5338-5351. 23. Livingstone M. S. and Hubel D. H. (1984) Anatomy and physiology of a color system in the primate visual cortex. J. Neurosci. 4, 309-357. 24. Merle P., Jarausch J., Trapp M., Scherka R. and Kadenbach B. (1981) Immunological and chemical characterization of rat liver cytochrome c oxidase. Biochim. biophys. Acta 669, 222-230. 25. Mita S., Schmidt B., Schon E. A., DiMauro S. and Bonilla E. (1989) Detection of "deleted" mitochondrial genomes in cytochrome-c oxidase-deficient muscle fibers of a patient with Kearns-Sayre syndrome. Proc. natn. Acad. Sci. U.S.A. 86, 9509-9513. 26. Montecucco C., Schiavo G. and Bisson R. (1986) ATP binding to bovine heart cytochrome c oxidase. Biochem. J. 234, 241-243. 27. Patterson T. E., Trueblood C. E., Wright R. M. and Poyton R. O. (1987) Polypeptide subunits encoded by nuclear genes are essential components of eukaryotic cytochrome c oxidase. In Cytochrome Systems: Molecular Biology and Bioenergetics (eds Papa S., Chance B. and Ernster L.), pp. 253-260. Plenum Press, New York. 28. Roberts H. and Hess B. (1977) Kinetics of cytochrome c oxidase from yeast membrane-facilitated electrostatic binding of cytochrome c showing a specific interaction with cytochrome c oxidase and inhibition by ATP. Biochim. biophys. Acta 462, 215-234. 29. Schatz G. and Mason T. L. (1974) The biosynthesis of mitochondrial proteins. A. Rev. Biochem. 43, 51-87. 30. Slonimski P. P. and Tzagoloff A. (1976) Localization in mitochondrial DNA of mutations expressed in a deficiency of cytochrome oxidase and/or coenzyme QH2-cytochrome c reductase. Eur. J. Biochem. 61, 27-41. 31. Sokoloff L. (1984) Metabolic Probes of Central Nervous System Activity in Experimental Animals and Man. Sinauer, Sunderland, MA. 32. Suske G., Enders C., Schlerf A. and Kadenbach B. (1988) Organization and nucleotide sequence of two chromosomal genes for rat cytochrome c oxidase subunit VIc: a structural and a processed gene. DNA 7, 163-171. 33. Tootell R. B. H., Silverman M. S., Hamilton R. L., Valois R. L. and Switkes E. (1988) Functional anatomy of macaque striate cortex. III. Color. J. Neurosci. 8, 1569-1593. 34. Trueblood C. E. and Poyton R. O. (1987) Differential effectiveness of yeast cytochrome c oxidase subunit V genes results from differences in expression not function. Molec. cell. Biol. 7, 3520-3536. 35. Virbasius J. V. and Scarpullar R. (1990) The rat cytochrome c oxidase subunit IV gene family: tissue-specific and hormonal differences in subunit IV and cytochrome c mRNA expression. Nucleic Acids Res. 18, 6581-6586~ 36. Wiesner R. J., Kurowski T. T. and Radovan Z. (1992) Regulation by thyroid hormone of nuclear and mitochondrial genes encoding subunits of cytochrome-c oxidase in rat liver and skeletal muscle. Molec. Endocr. 6, 1458-1467. 37. Wong-Riley M. T. T. (1979) Changes in the visual system of monocularly sutured or enucleated cats demonstrable with cytochrome oxidase histochemistry. Brain Res. 171, 11-28. 38. Wong-Riley M. T. T. (1989) Cytochrome oxidase: an endogenous metabolic marker for neuronal activity. Trends Neurosci. 12, 94-101. 39. Wong-Riley M. and Carroll E. W. (1984) Effect of impulse blockade on cytochrome oxidase activity in monkey visual system. Nature 307, 262-264. 40. Wong-Riley M. T. T., Merzenich M. M. and Leake P. A. (1978) Changes in endogenous enzymatic reactivity to DAB induced by neuronal inactivity. Brain Res. 141, 185-192. 41. Wong-Riley M. T. T., Walsh S. M., Leake-Jones P. A. and Merzenich M. M. (1981) Maintenance of neuronal activity by electrical stimulation of unilaterally deafened cats demonstrable with the cytochrome oxidase technique. Ann. Otol. Rhinol. Lar. 90, Suppl. 82, 30-32. 42. Wong-Riley. M. T. T., Tripathi S. C., Trusk T. C. and Hoppe D. A. (1989) Effect of retinal impulses blockage on cytochrome oxidase-rich zones in the macaque striate cortex: I. Quantitative electron-microscopic (EM) analysis of neurons. Vis. Neurosci. 2, 483-497.

210

S. Liu and M. Wong-Riley

43. Wong-Riley M. T. T., Trusk T. C., Tripathi S. C. and Hoppe D. A. (1989) Effect of retinal impulse blockage on cytochrome oxidase-rich zones in the macaque striate cortex: II. Quantitative electron-microscopic (EM) analysis of neuropil. Vis. Neurosci. 2, 499-514. 44. Wright R. M., Dircks L. K. and Poyton R. O. (1986) Characterization of COX 9, the nuclear gene encoding the yeast mitochondrial protein cytochrome c oxidase subunit VIIa. J. biol. Chem. 261, 17183-17191. 45. Yamada M., Amuro N., Goto Y. and Okazaki T. (1990) Structural organization of the rat cytochrome c oxidase subunit IV gene. J. biol. Chem. 265, 7687-7692. (Accepted 9 January 1994)