Subcellular compartmentalization of CD38 in non hematopoietic cells a study to characterize its functional role in mitochondria 7

4.3 Discussion Although the possible presence of functional CD38 in the mitochondria was previously shown using specific organelle targeting CD38 expression system in cell lines Chapter

4.2.5.2 Localization of CD38 on Percoll purified mitochondria using Scanning Electron Microscopy (SEM)

The results obtained in the localization study of CD38 on mitochondria extracted from mouse brain tissues, labeled with 15nm colloidal gold marker, after mild prefixation, are illustrated in Figure 4.21-4.23 CD38 labeling was observed in the BEI (Backscatter electron imaging) mode The contrast is dependent on the atomic number, and the gold marker (Au; Z=79) is rendered in high contrast, as well

as presenting a clearly distinguishable matrix structure of mitochondria The gold particles were well visualized in the BEI mode because the 20kV electron beam penetrates most surface structures and good contrast It is well established that cell surface labeling with SEM is useful as direct correlation between the presence of specific molecules exposed on cell surfaces (antigen, receptor sites, etc) and the

surface structure of these same cells (de Harven et al., 1984) Thus the comparison of

the SEM (Figures 4.21-4.23) and TEM (Figure 4.18) with gold labeling of CD38 on mitochondria showed that the localization of the molecule is restricted to the cell surface

It is noted that the number of gold particles per surface area is markedly higher in the WT mouse brain mitochondria as compare to the KO mouse brain mitochondria (Figure 4.22) This is interpreted as reflecting the high specificity of the CD38 antibody used which showed negligible background or non-specific staining (Figure 4.22 & 4.23) The inset shows the enlargement of the encircled area, B, of the isolated WT mouse brain mitochondria (Figure 4.22 B) as well as the isolated CD38

KO mouse brain mitochondria (Figure 4.23 A-C) The encircled areas in Figure 4.22, A-F were observed with intense distribution of gold particles which appeared white in BEI mode In contrast, CD38 KO mouse brain mitochondria in Figure 4.23 shows

negative staining, as indicated by negligible observation of gold particles as compared

to the WT A closer look at the samples by higher magnification (Figure 4.23 A-C) supported the observations of gold particle staining on the WT sample There was negligible staining for the CD38 KO samples

The same portion of the same sample was illustrated by superimposition of both the BEI and the SEI (Secondary Electron Imaging) signals (signal mixing

(Becker et al., 1979) The polarity of the backscatter signal was set back to ‘normal’

and the SEI and BEI signals were overlaid The result was an image in which the distribution of all the gold particles can be precisely correlated with details of the surface structure of the sample It was observed that the majority of discernible gold markers were associated on the surface of the purified intact mitochondria (Figure 4.21, 4.22) Collectively, the present data further supported the outer mitochondrial location of CD38 and strongly agreed to the specific topology with protrusion of C-terminal domain outside the mitochondria, which has consistently observed in the earlier experiments (Chapter 3), as well as the TEM, digitonin titration and protease protection assay

Figure 4.22 Backscatter electron imaging of mitochondrial fractions, labeled with

goat polyclonal CD38 antibody-sc-7049 Backscatter electron micrograph at low magnification with the labeling of CD38 on purified mitochondria isolated form WT mice brain tissues The encircled areas (A-F) are CD38 immunoreactive region that detected by conjugated gold particles The inset is enlargement of the encircled area (B) of the isolated mitochondria

Scale bar: 1µm

B

A

E D

C

B

F

B

Figure 4.23 Backscatter electron imaging of CD38 KO mitochondrial fractions, labeled with goat polyclonal CD38 antibody-sc-7049 Backscatter electron micrograph at low magnification with the labeling of CD38 on purified mitochondria isolated form CD38KO mice brain tissues The encircled areas (A-C, please refer to next page) were examined at higher magnification for positive staining of CD38 The insets show enlargement of the encircled area (A-C shown in the following page) with negligible CD38 staining

Scale bar: 1µm

C

A

B

B A

4.3 Discussion

Although the possible presence of functional CD38 in the mitochondria was previously shown using specific organelle targeting CD38 expression system in cell lines (Chapter 3), this is the first conclusive evidence of the presence of CD38, a hitherto characterized multifunctional enzyme, on the mitochondria membrane via an endogenous system, namely highly purified mitochondria isolated from mouse brain tissues In addition to the finding of intracellular CD38 localized on nucleus, ER

(Adebanjo et al., 1999; Khoo et al., 2000; Sun et al., 2002), the results described in

this study provide an alternative resolution to the topological issue of this molecule

Studies on NAD+ glycohydrolase activities in mitochondria started back in the

1980s Moser et al (1983), Masmoudi et al (1987;1988) and Hilz et al (1984) have

independently reported an unidentified NAD+ glycohydrolase activities in mitochondria isolated from rat liver, brain and heart tissues, respectively All known observations of mitochondria NAD+ glycohydrolase showed high activity with NAD+, poly-(ADP-ribose) polymerase activity was not detected in the specific

enzyme and it was susceptible to inhibition by nicotinamide and ATP (Moser et al., 1983; Masmoudi et al., 1988) The fact that this NAD+ glycohydrolase observed in the brain mitochondrial preparation was identified as likely to be CD38 in the present study (Table 4.1) suggests that the subcellular distribution of this molecule is more complex than originally thought

More information about the mitochondrial NAD+ glycohydrolase was revealed

in the next two decades Ziegler et al (1997; 1997b) and Liang et al (1999) reported

independently of mitochondrial NAD+ glycohydrolase activity observed in bovine liver and rat liver tissues Both Ziegler and Liang groups’ investigations identified the liver mitochondrial NAD+ glycohydrolase as a member of the class of bifunctional

ADP-ribosyl cyclases/cyclic ADP-ribose hydrolases, which is involved in synthesis and degradation of cADPR, the well characterized potent intracellular calcium-mobilizing agent NAADP was first shown produced in rat liver mitochondrial NAD+

glycohydrolase by Liang et al (1999) Ziegler group reported that the bovine

mitochondrial NAD+ glycohydrolase showed comparable yield of cADPR and ADPR

to CD38 It was also found that Zn2+ increased the yield of cADPR by more than 3 fold in bovine NAD+ glycohydrolase (Zocchi et al., 1993; Ziegler et al., 1997) This

result was contradicted by Liang group which showed otherwise The properties of NAD+ glycohydrolase from a given tissue may vary widely with the species and,

within a species, may vary from tissue to tissue (Green et al., 1964) This may

account for the discrepancy Interestingly, Liang group detected a minimal quantity of CD38-like antigen in liver mitochondria

Lisa et al (2001) reported that 92% of NAD+ glycohydrolase activity in rat

heart was found in mitochondria Aksoy et al (2006) has provided substantial

evidence that gave support to the novel concept of CD38, a widely recognized NAD+ glycohydrolase as the major regulator of intracellular NAD+ in several tissues including brain In agreement to this, the total mitochondria fraction isolated from mouse brain tissues in the present study showed significantly higher NAD+ glycohydrolase activity as compared to ADP-ribosyl cyclase activity (Chapter 4 Section 4.2.4.2) No apparent cyclase and NAD+ glycohydrolase activities were observed in CD38KO mouse brain mitochondria samples (Figure 4.9), which is in

agreement with the finding reported in Aksoy et al., 2006 Hence, the present data

demonstrating presence of CD38 on mitochondria is well supported and eliminates

the possibility of contamination by unidentified ADP-ribosyl cyclase (Ceni et al.,

2003b; 2006) known to be found in CD38KO mouse brain tissues

To learn about the cellular function of this mitochondrial CD38, it is important to examine its precise localization on the organelle Mitochondria are compartmentalized into matrix, inner mitochondria membrane, intermembrane space and outer mitochondrial membrane, as discussed in Chapter 3 The outer membrane has different characteristics including different cholesterol composition as well as different protein composition from the inner mitochondrial membrane Both membranes come into close contact showing membrane contact points, which are important in regulation of the mitochondria-cytoplasmic exchanges of metabolites, proteins, and phospholipids (Brdiczka, 1991)

Recent data argues that mitochondrial NAD+ glycohydrolase is localized to

the outer mitochondrial membrane (Boyer et al., 1993; Yamada et al., 1997; Lisa et

al., 2001) in contrast to the previous believes that mitochondrial NAD+ glycohydrolase is localized to the inner side of the inner mitochondrial membrane

(Lotscher et al., 1980; Moser et al., 1983) It was well established that a loss of

intramitochondrial pyridine nucleotides occurs during prooxidant mediated Ca2+ efflux from Ca2+ loaded mitochondrial While the inner membrane of mammalian mitochondria is normally impermeable to pyridine nucleotides, it was shown that the pyridine nucleotides could be transported through the inner mitochondrial membrane

via the opening of the permeability transition pore (PTP) (Boyer et al., 1993; Lisa et

al., 2001) Lisa et al further showed that the redistribution of NAD+ between these 2 compartments is made possible by PTP opening because mitochondrial swelling precedes NAD+ hydrolysis mediated by NAD+ glycohydrolase Thus the pyridine nucleotides are transported to the outer mitochondrial membrane where hydrolysis takes place

In agreement with the published data mentioned above, the results obtained from digitonin titration showed a complete solubilisation of CD38 molecule at ≤ 0.1mg/mg of digitonin to protein ratio, indicated that the mitochondrial location of CD38 co-localizes with Tom20 as well as Bcl-xL (the known outer mitochondrial membrane markers) suggesting that its location is the outer mitochondrial membrane (Figure 4.7) This is further supported by the integrity of the intermembrane space proteins and inner mitochondrial membrane proteins under the same treatment

More evidence gathered when protease protection assay was introduced into the digitonin solubilization system offered further confirmation of the location of CD38 to the outer mitochondrial as well as the specific topology in which its carboxyl catalytic domain extruding to the cytosolic region These pieces of evidence include 1) CD38, as well as the outer mitochondrial membrane markers, Tom20 and Bcl-xL were susceptible to protease digestion even before subjected to digitonin solubilisation; 2) The CD38 antibody, sc7049 (M19, Santa Cruz) employed here detects epitope mapped to extracellular domain of the molecule However, the nature

of the study was unable to reach definitive conclusions with regards to the precise topography of CD38 on the outer mitochondrial membrane This is due to the limitation of choices with immunoreactive mouse CD38 antibody The protein topology can be further investigated by means of peptide antibodies raised against the carboxyl-terminal region of the protein

Endoplasmic reticulum/microsomal vesicles have been shown to contain

CD38 molecule (Sun et al., 2002), and mitochondrial preparation are commonly contaminated with varying amounts of microsomes (Schnaitman et al., 1968) From

the data presented (Figure 4.6), isolated purified mitochondria was shown to be devoid of ER as well as plasma membrane markers Thus the results of this study

indicate that CD38 observed in Percoll purified mouse brain mitochondrial preparations is a native constituent of the outer mitochondrial membrane and that the outer membrane is the only mitochondrial location for this molecule

The results obtained from digitonin titration and protease protection assay were further verified by employing electron microscopic techniques The localization

of CD38 as well as the topography on the mitochondrial was examined by combining TEM and SEM methods The TEM consistently showed that the distribution of DAB staining was strictly localized to the outer rims of mitochondria, as would be expected for protein associated in the outer mitochondrial membrane This is generally in

agreement with those reported by Yamada et al (1997) Three CD38 antibodies were

employed in the immunostaining experiment, two polyclonal CD38 commercialised antibodies and a lab customized peptide anti-human CD38 antibody All three antibodies recognized CD38 and showed the same staining pattern This is not surprising as it is well established that murine CD38 cDNA shows 70% sequence

homology with that of human CD38 (Harada et al., 1993) Consistent labelling of

intracellular CD38 on nuclear envelope was not observed; in fact the intracellular CD38 staining was highly heterogeneous but remained constant for plasma membrane and postsynaptic densities The difference could be due to the use of mouse brain tissues in current study as opposed to use of rat brain tissues in Yamada’s data It is noteworthy that the staining pattern on mitochondria is specific and heterogeneous; uneven rather than uniform distribution of CD38 associated with the outer mitochondrial membrane was observed Not all mitochondria in the immunopositive cells were stained and not all cells are immunoreactive to CD38 (Figure 4.10-4.17) The observation is interesting, in account of accumulative research in the connectivity and functional homogeneity of mitochondria in cells; recent publications showed the