Murine Spinal Cord Transcriptome Analysis Following Systematic Reduction of Prevalent Myelin cDNA Sequences

INTRODUCTION

Using suppression subtraction hybridization (SSH) (Diatchenko et al. 1996) and mirror orientation selection (MOS) (Rebrikov et al. 2000), we have previously demonstrated that a number of genes are differentially expressed in murine spinal cord without sensory ganglia when subtracted against the visual cortex, a motor neuron poor area of the brain (Lathia et al, 2006).� A large percentage of these differentially expressed genes (44 percent) are of known glial origin; primarily proteolipid protein (Plp1), ferretin heavy chain (Fth1) and myelin basic protein (Mbp) among others. Similar results are reflected in the expression profiles for many myelin genes in the murine spinal cord in the expressed sequence tag database (dbEST) held at the NCBI UniGene project (Pontius et al. 2003).�

Subtraction of any brain area against any other brain area including the spinal cord would necessarily include a large amount of myelin which should subtract out as a common element.� It seems unlikely, therefore,� that glial sequences are being expressed in the spinal cord merely as a result of an overabundance of glia in this structure.� The amount of various myelin proteins; proteolipid protein (PLP), myelin basic protein (MBP) and 2�, 3� cyclic nucleotide 3� phosphodiesterase (CNP) are not increased in spinal cord over that of other brain areas in human tissue (Trotter et al. 1984), however, the prospect that glial proteins are continuing to be produced in the spinal cord after significant turnover has slowed in other areas of the nervous system cannot be ruled out.

It is possible that our results, as well as previous attempts by others to demonstrate differences between areas in the nervous system (Kobayashi et al 1991; Usui et al. 2003; Akopian and Wood, 1995; de Chald�e et al. 2003 and Sandberg et al. 2000), record small differences between these structures because of the overwhelming presence or activity of glial elements co-expressing and obscuring the less abundant neural elements.�

In an attempt to reduce the presence of glial derived sequences, we used spinal cord as the tester and isolated white matter as driver and vice versa� in the present study in order to examine the spinal cord transcriptome.� A continued presence of a significant number of glial derived genes differentially expressed in the spinal cord would indicate that myelination or other glial activities are increased over other brain areas tested in the adult mouse.

MATERIALS AND METHODS

Tissue Preparation

Tissues for all subtractions were from 90-110 day old male C57Bl6J mice housed at the Jackson Laboratories in Bar Harbor, Maine.� The animals were sacrificed by cervical dislocation and in all cases the brains immediately removed and stripped of meninges.� Spinal cord central cores were dissected out and white matter and corpus callosae were manually dissected away from the remainder of the brain.� All dissected tissues were immediately frozen on blocks of dry ice.� Tissues were shipped to Solomon Park on dry ice and held at -70 degrees C. until ready for use.

RNA Isolation

Total RNA was isolated by the method of Chomczynski & Sacchi (1987).� Messenger RNA was further isolated on poly T columns purchased from Ambion.

Suppression Subtractive Hybridization

Suppression subtractive hybridization was performed using The BD Clontech PCR-Select� cDNA Subtraction Kit with cDNA from spinal cord as tester and cDNA from white matter as driver or vice versa in reverse subtractions.� cDNA from each specimen was prepared from 2 �g Poly (A)+.� RNA preparations were digested with the endonuclease RSA 1.� Tester cDNA were then divided into 2 subpopulations and each ligated with either adaptor 1 or adaptor 2R from Clontech.� Subtractions and normalizations were accomplished with a 50 fold excess of driver over tester followed by 2 rounds of PCR amplification.� Subtraction efficiency was measured by the intensity of the PCR products on agarose gel representing glucose 3 phosphodiesterase before and after the SSH procedure.

Transformation

Secondary PCR products were transformed into pCR� 4-TOPO vector with TOP10 E. coli cells (InVitrogen).

Probe preparation

Approximately twenty five nanograms of secondary PCR products from forward and reverse subtractions were labeled using 50uCi of [32p] dCTP with Rediprime II Random Prime Labeling System from Amersham. Labeled probes were purified using BD Chroma Spin Columns from BD Biosciences

Differential Screening

Differential screening was performed by dot blot analysis using the PCR-Select Differential Screening Kit from BD Biosciences.� Identical sets of subtracted clones were spotted onto four nylon transfer membranes (Schleicher and Schuell). The membranes were subsequently interrogated with [32P] labeled cDNA (RediprimeII random prime labeling system from Amersham Biosciences) from the forward and reverse subtracted and unsubtracted probes.� Clones exhibiting at least two fold differences only in the forward subtraction were included in the subtracted library.

Sequencing

Sequences from secondary PCR products were obtained from Retrogen in San Diego, CA.

Northern Hybridization Procedure

Total RNA (50�g) from murine spinal cord, or white matter was separated by gel electrophoresis on 1% agarose � formaldehyde gel using the northernMax Kit from Ambion and subsequently blotted onto a nylon hybridization membrane from Ambion.�� Custom primers were synthesized and subsequently, amplified and purified from low melting gels using Qiaquick gel purification kit from Qiagen.

Comparison of Current Results with UniGene Records

All genes exhibiting higher expression in the spinal cord were compared to the UniGene entries listed at the NCBI open access data base for Mus musculus at (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=unigene).� All sequences were scored with the estimated number of sequences per million.� Sequences marked with a double dagger (�) represent sequences that were the top three most frequent in the spinal cord for these genes.� Sequences that were not found in the murine spinal cord in UniGene even though measured were marked as zero.� Sequences not previously measured in the murine spinal cord as well as ESTs never before reported are marked in the tables as [no data].� UniGene data were current as of March 31, 2007.

RESULTS

Overall Analysis of Differentially Expressed Sequences

Approximately 1,000 colonies across 3 separate subtractions of spinal cord subtracted against white matter, were randomly selected for analysis.� Two hundred and twenty clones from these colonies were identified by dot blot analysis as differentially expressed in spinal cord and subsequently sequenced and identified using the basic local alignment and search tool (BLAST) programming available through the NCBI.� From these, 140 unique genes were identified.� The results of these subtractions are summarized in table I (A-F) along with the current estimate of sequences per million for each identified EST in the UniGene database. Protein products were classified from either Entrez Online Mendelian Inheritance in Man (OMIM) at (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=OMIM), Entrez Gene at (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene ), Entrez Protein at (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=Protein) or from Pubmed at (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed ).

See Table I A-F

Dot Blot Analysis of Differentially Expressed Sequences

A representative dot blot analysis is shown in figure 1.� Approximately 50 percent of the cloned sequences show increased labeling.� The labeled samples were scored and only those represented by at least a two fold increase in labeling intensity when tested by the forward subtracted probe versus forward unsubtracted and reverse subtracted and unsubtracted probes were selected for inclusion in the subtracted library.

See Figure 1

Northern Analysis of Differentially Expressed Sequences

All sequences used in the northern analyses were selected from dot blot confirmed, differentially hybridized libraries.� These northern analyses are shown in figure 2.� All colonies used were randomly drawn from one of three categories from the UniGene database held at the NCBI.� The first category represents genes that are not yet documented in murine spinal cord: secreted phosphoprotein 1-osteopontin (Spp1), myelin protein zero (Mpz) and RIKEN cDNA 111000P04 (111000P04Rik).� The second category is moderately represented in spinal cord (RIKEN cDNA 4732416N19 (4732416N19Rik), peripheral myelin protein 22 (Pmp22), flotillin 2 (Flot2) and ATPase, Na+/K+ transporting alpha 2 polypeptide (Atp1a2).� The third category is from genes that are highly represented in spinal cord: Glial fibrillary acidic protein (Gfap), kallikrein 6 (Klk6), neurofilament heavy chain (Nefh) and neurofilament medium chain (Nefm). The increase in labeling for all genes is substantial with the exception of Flot2 which shows an approximately 50 percent increase by density scanning (data not shown) in spinal cord versus white matter.�� The Northern blots representing these sequences were normalized with β actin as shown in the figure.

See Figure 2

Differential Expression of Myelin Associated Sequences

Eighteen clones representing nine known, unique myelin associated sequences were differentially expressed in spinal cord when subtracted against white matter and represent 10 percent of all clones in this subtraction as represented in Table II.�� Entries marked with the double dagger (�) are expressed among the top three most frequently represented in spinal cord in the UniGene database.

The number of clones represented by all known glial cell derived sequences is reduced from 44 percent when spinal cord was subtracted against visual cortex in our previous study to 10 percent when this structure is subtracted against isolated white matter.� The number of clones represented by Plp1 was

reduced from 22 percent in our previous study (spinal cord subtracted against visual cortex) to less than one percent in the present study.�

See Table II

White Matter Subtracted Against Spinal Cord

Three hundred colonies across two separate subtractions where myelin was subtracted against spinal cord (reverse subtraction) were tested for differential expression by dot blot analysis.� Out of 56 differentially expressed clones, 28 were sequenced and of these, 18 were unique and are listed in table III.� The genes found to be differentially expressed in isolated brain white matter when this material was subtracted against spinal cord contained no known myelin specific genes.

See Table III

Novel Expressed Sequences not Represented by mRNA Sequences

Four of the ESTs in the present study differentially expressed in spinal cord (NCBI accession numbers ES387449, ES387454, ES387457 and ES387458) are novel and have been recorded in the NCBI� database.��

Expressed Sequence not previously measured in murine spinal cord

One of the ESTs in the present study, suppressor of lin-12 1 homolog (C. elegans), Sel1h, although adequately represented with numerous ESTs has not been measured in the mouse spinal cord to date and therefore has no entry in the UniGene database for this structure.��

DISCUSSION

General Considerations

The complexity of the transcriptome of any nervous system is enormous.� This is especially evident in mammalian nervous systems where the number and variety of neurons is most probably at least in the hundreds if not the thousands, each contributing an array of messenger RNA molecules to the overall transcription occurring in the system at any particular time (Stevens 1998; MacNeil et al.1999).� Adding to the difficulty of obtaining any meaningful information from individual neurons in such a system is the large number of glial cells outnumbering neurons by at least ten and possibly a hundred fold in some areas.

Various techniques have been developed over the past several years that allow the transcriptomes of either individual cells or tissues or whole organs to be analyzed for mRNA presence and abundance.� The most widely used of these methods are laser capture (Emmert-Buck et al., 1996) for the isolation of mRNA from individual cells, (however see Kodha et al. 2000), microarray technology (Schena et al. 1995); serial analysis of gene expression (SAGE) (Velculescu et al. 1995) and subtractive hybridization (Hedrick et al. 1984) with a number of other techniques derived from these basic procedures for populations of cells.� Suppression subtractive hybridization (SSH) alone simultaneously allows both the direct comparison of two tissues and the detection of rare or low abundance as well as unknown transcripts (Shusta, 2005).

By subtracting spinal cord directly against isolated white matter, we have eliminated nearly three quarters of the known glial sequences previously documented by us in subtractions against visual cortex containing underlying white matter.� Using isolated myelin as driver allows us to identify an array of differentially expressed genes more consistent with the neural elements making up this structure.

Comparison to UniGene Results

�� There is a good correlation with our approach to the data already accumulated from many sources for the spinal cord as represented by the UniGene dbEST expression profiles.� Approximately one in every three of our clones (49 out of 140) is represented by sequences that are among the highest in frequency in the existing dbEST data for spinal cord. More significantly, nearly an equal number (45 out of 140), although sought, have never been detected in the spinal cord as represented by the same database.�� Eight percent (4 out of 211) are new ESTs that have not been demonstrated in the mouse to date.� In total, 49 of the 140 differentially expressed genes or 35% have not been shown to be expressed in the spinal cord to date indicating that a large percentage of genes differentially expressed in this structure are unreported.

Specific Observations concerning Differentially Expressed genes

The spinal cord subtracted against isolated myelin shows many differentially expressed sequences (24 per cent) representing proteins concerned with structural and motor activities involved with maintaining and transporting molecules and organelles along the extended axons known to be present in this structure.� The next largest portion (27 per cent) are sequences representing enzymes, a large portion of these (8 per cent of the total number of differentially expressed genes) are concerned with either glycolysis or lipid metabolism.�� A number of genes representing ion channels and other uptake functions are well represented as might be expected for the spinal cord neural population.

The strong showing of Spp1 (osteopontin) as well as Sparc (osteonectin) deserves some mention.� We have also observed these two genes differentially expressed when we subtracted spinal cord against visual cortex (Lathia et. al., 2006).� Neither of these genes is recorded as differentially expressed in the spinal cord in the UniGene database.� One explanation for these genes being differentially expressed is their correlation with growth factors such as connective tissue growth factor (Denhardt & Noda 1998, Rittling & Denhardt 1999) which again should be expected in the spinal cord.�

Leucyl t-RNA synthetase (Lars2) is strongly differentially expressed in the spinal cord.� This is a verification of our previous subtractions of spinal cord against visual cortex.�� In the present study, Lars2 represents 12 percent of the total clones differentially expressed and was eight percent of the total number of clones when subtracted against visual cortex.� No other aminoacyl t-RNA synthetase is represented in either subtraction.� The aminoacyl t-RNA synthetases are known to be phylogenetically old and as such it is speculated that a number of these proteins may have evolved alternate roles (Martinis, S.A. et al., 1999).� It is known, for example, that human cytoplasmic tRNA synthetases are assembled into large macromolecular complexes during protein synthesis which masks any alternate functions concerned with the immune response, apoptosis and angiogenesis (Wakasugi et al. 1999 and Wakasugi et al. 2002).� In microorganisms, leucyl-tRNA synthetases have been show to facilitate RNA splicing activities (Houman et al. 2000). Finally, Lars2 also has been shown to have a critical amino acid proofreading activity to ensure the fidelity of protein synthesis (Karkhanis et al. 2006).

When we reversed our subtraction and subtracted white matter from spinal cord, we found no known myelin or glial associated genes differentially expressed, strongly indicating that the presence of differentially expressed sequences in the spinal cord represent genes that are truly differentially expressed and not simply the product of an excessive amount of glial derived sequences.�

At the age that the animals were tested (90-110 days), it is generally assumed that the formation of myelin has been completed.� The persistence of genes associated with myelin in adult murine spinal cord when subtracted against white matter can best be explained by assuming that the spinal cord is simply exhibiting higher rates of myelin production or metabolism than the other brain areas measured at this age.� The driver cDNA, in this case the underlying white matter from all areas of the cerebral cortex, presents sufficient cDNA from myelin to subtract out any cDNA from glial cells in the spinal cord.

Our results also confirm that SSH represents the best approach to elucidating the transcriptomes of nerve cells isolated from the overwhelming influence of the surrounding glial cells.� The large number of genes differentially expressed in our study which are recorded as not present in the spinal cord by the large dbEST indicates that a significant percentage (nearly a third) of differentially expressed genes are being unreported by most studies.�

By selecting the appropriate driver and tester in the subtraction not only can the background of myelin be greatly reduced or eliminated, but sub classes of neurons can be subtracted against each other eventually allowing a single cell type to be analyzed.� The only viable alternative to this would be an isolation of single cells directly and although technically possible this procedure presents problems not only in the length of time that is required to fix and mechanically isolate the cells, but also in that the amount of RNA ultimately achieved is small and the amplification of these sequences is known to introduce significant artifacts into the final analysis (Nygaard et al. 2005).� SSH on the other hand represents a rapid and sensitive procedure for quantifying transcription events in the nervous system, however, it should be noted that all procedures necessarily take some time to extract and process the tissues of interest leaving the possibility that some genes may be differentially expressed due to extraneous variables.

Finally, the persistence of myelin associated sequences in the spinal cord raises the possibility that diseases that preferentially affect this structure, such as motor neuron disease, can do so through these cells.� It is obvious that the extended structural demands of motor neurons as well as their enormous energy requirements and the known uptake and processing of neurotransmitters by surrounding cells, make the nerve� cells served by these glial cells in the spinal cord potentially vulnerable to many insults from within.�

CONCLUSION

�� When adult, murine spinal cord is subtracted directly against isolated white matter and corpus callosum, the number of myelin associated genes differentially expressed in this structure is reduced but not eliminated from that seen in previous studies where spinal cord was subtracted against visual cortex with underlying white matter.� Myelination or myelin associated activity is clearly differentially expressed in spinal cord, complicating attempts to view the transcriptomes of neural elements in this structure.� Comparison of these results with established expression databases indicates that large numbers of genes are not being reported.

ACKNOWLEDGMENTS

�� This work is supported by a bequest from the estate of Frances and Wesley Johnson and the generous donations of numerous individuals to Solomon Park Research Institute.� We would like to thank the members of the board of directors of Solomon Park Research Institute for their tireless dedication to the Institute and the ongoing research effort.

�� We are especially appreciative of the many helpful remarks and suggestions from Dr. Susan A. Martinis concerning the role of leucyl-tRNA synthetase.

REFERENCES

Akopian, A.N., and Wood, J.N. Peripheral nervous system-specific genes identified by subtractive cDNA cloning.� JBC (1995) 270 (36) 21264-21270.

Chomczynski, P., and Sacchi, N. Single-step Method of RNA Isolation by Acid Guanidinium Thiocyonate-Phenol-Chloroform Extraction, Anal. Biochem. (1987) 162, 156.

De Chaldee, M., Gaillard, M-C., Bizat, N., Buhler, J-M., Manzoni, O., Bockaert, J., Hantraye, P., Brouillet, E. and Elalouf, J-M. Quantative assessment of transcriptome differences between brain territories. Genome Research (2003) 13 1646-1653.

Denhardt, D.T. and Noda, M. Osteopontin expression and function:

role in bone remodeling. Journal of Cellular Biochemistry (1998) 30�31

(Suppl) 92�102.

Diatchenko L, Lau Y, Campbell A, Chenchik A, Moqadam F, Huang B, Lukyanov S, Lukyanov K, Gurskaya N, Sverdlov E, Siebert P (1996). Suppression subtractive hybridization: a method for generating differentially regulated or tissue-specific cDNA probes and libraries. Proc Natl Acad Sci U S A 93 (12): 6025-30.

Emmert-Buck, M.R.; Bonner, R.F.; Smith, P.D.; Chuaqui, R.F.; Zhuang, Z.; Goldstein, S.R.; Weiss, R.A. and Liotta, L.A. (1996)� Laser Capture Microdissection Science 274(5289):998-1001

Hedrick, S.M., Chen, D.I., Nielsen, E.A., and Davis, M.M.� Isolation of cDNA clones encoding T cell-specific membrane-associated proteins. Nature (1984)308, 149.

Houman, F., Rho, S.B., Zhang, J., Shen, X., Wang, C.-C., Schimmel, P., and Martinis, S.A. (2000) A Prokaryotic and Human tRNA Synthetase Provide an Essential RNA Splicing Function in Yeast Mitochondria. Proc. Natl. Acad. Sci., U.S.A., 97, 13743-13748.

Karkhanis, V.A., Boniecki, M.T., Poruri, K ., and Martinis, S.A. (2006) A Viable Amino Acid Editing Activity in the Leucyl-tRNA Synthetase CP1 Splicing Domain is not Required in the Yeast Mitochondria. J. Biol. Chem., 281, 33217-33225.).

Kobayashi, H., Takahashi, H., Oyanagi, K., Ikuta, F., Miyatake, T., and Tsuji, S. Construction of Spinal Cord cDNA library and application for subtractive cloning of spinal cord-specific cDNAs. J. Mol. Neurosci. (1991) 3:59-64.

Kohda, Y.; Murakami, H.; Moe, O.W. and Star, R.A. (2000) Analysis of segmental renal gene expression by laser capture microdissection. Kidney International Vol. 57 pp 321-331

Lathia, K.B., Yan, Z., and Clapshaw, P.A. (2006) Spinal Cord Transcriptome Analysis Using Suppression Subtractive Hybridization and Mirror Orientation Selection. Cell. Mol. Neurobiol. 26(3):259-75.

MacNeil, M.A. and Masland, R.H. (1998) Extreme diversity among amacrine cells: Implications for function. Neuron 20:971-982.

Martinis, S.A., Plateau, P., Cavarelli, J., and Florentz, C. (1999) Aminoacyl-tRNA Synthetases: An Expanding Family of Functions. EMBO J., 18, 4591-4596

Nygaard, V., Holden, M., L�land, A., Langaas, M., Myklebost, O. & Hovig, E. (2005) Limitations of mRNA amplification from small-size cell samples. BMC Genomics 6:147 Published online 2005 October 27. doi: 10.1186/1471-2164-6-147.

Pontius, J.U., Wagner, L., Schuler, G.D. (2003) UniGene: a unified view of the transcriptome. In: The NCBI Handbook. Bethesda (MD): National Center for Biotechnology Information.

Rebrikov, D.V., Britanova, O.V., Gurskaya, N.G., Lukyanov, K.A., Tarabykin, V.S. and Lukyanov, S.A.. (2000) Mirror orientation selection (MOS) � a method for eliminating false positive clones from libraries generated by suppression subtractive hybridization. Nucleic Acid Research (20) 20:e90

Rittling, S.R.and Denhardt, D.T. Osteopontin function in pathology:

lessons from osteopontin-deficient mice. Experimental Nephrology (1999) 7

103�113.

Sandberg, R., Yasuda, R., Pankratz, D.G.; Carter, T.A., Del Rio, J.A. Wodicka, L., Mayford, M., Lockhart, D.J. and Barlow, C.� Regional and strain-specific gene expression mapping in the adult mouse brain.� Proc. Natl. Acad. Sci. USA (2000) 97 11038-11043.

Schena M, Shalon D, Davis RW, Brown PO. (1995). Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science. Oct 20; 270 (5235): 467-70.

Shusta, E.V. (2005) Blood-Brain Barrier Genomics, Proteomics, and New Transporter Discovery. NeuroRx Vol. 2, No. 1. pp 151-61.

Stevens C.F. (1998)Neuronal diversity: too many cell types for comfort? Curr Biol. 8(20):R708-10

Trotter, J.L., Wegescheide, C.L., and Garvey, W.F. (1984) Regional studies of myelin proteins in human brain and spinal cord. Neurochem. Res. 9(1):133-146

Usui, H., Falk, J.D., Dopazo, A., de Lecea, L., Erlander, M.G. and Sutcliffe, J.G., Isolation of clones of rat striatum-specific mRNAs by directional Tag PCR subtraction.� J. Neurosci. (1994) 14(8) 4915-4926.

Velculescu VE, Zhang L, Vogelstein B, and Kinzler KW. (1995) Serial Analysis of Gene Expression. Science 270:484-7

Wakasugi, K. and Schimmel, P. (1999). Two distinct cytokines released from a human aminoacyl tRNA synthetase. Science 284: 147-151.

Wakasugi, K., Slike, B., Hood, J., Otani, A., Ewalt, K. L., Friedlander, M., Cheresh, D. A., and Schimmel, P. (2002). A human aminoacyl-tRNA synthetase as a regulator of angiogenesis. Proc. Natl. Acad. Sci. USA 99: 173-177).

Huminiecki L, Lloyd AT, Wolfe KH. (2003)Congruence of tissue expression profiles from Gene Expression Atlas, SAGEmap, and TissueInfo databases. BMC Genomics 4:31.

Table I Gene Sequences Differentially Expressed in Spinal Cord versus White Matter
Table I -A *Structural and Motor Protein Sequences*
Gene Abbreviation	Gene Frequency	UniGene Frequency	Putative and Known Functions
Spp1	5	0	bone structure osteoclasts (osteopontin)
Mpz	3	0	myelin adhesion protein (Schwann Cells)
Myo5a	2	0	motor activity (centriole)
Elmod2	1	0	cytoskeleton
Mast1	1	0	motor microtubule activity
Tnni2	1	0	muscle contraction
Ttn	1	0	muscle structure and function
Anln	1	37	cytoskeleton (actin binding)
Nebl	1	37	cytoskeleton (actin binding)
Sparc	2	75	bone structure osteoclasts (osteonectin)
Kif5c	1	75	motor axon trafficking
Scg3	1	75	muscle actin binding sarcolemma
Ankfy1	1	75	trafficking vesicle
Sept8	1	75	trafficking vesicle/neuronal polarity
Rer1	1	75	trafficking Golgi
Kif5b	2	150	motor microtubule activity
Ap3m1	1	150	clathrin binding
Dynll2	1	187	motor microtubule activity
Gfap	3	112�	myelin intermediate filaments (oligodendroglia)
Nefh	4	187�	cytoskeleton (neural)
Ap1m1	2	225�	clathrin binding
Lrpprc	1	225�	cytoskeleton with vesicular trafficking
Nefm	4	300�	cytoskeleton (neural)
Plp1	2	712�	myelin sheath
Mbp	1	2026�	myelin sheath
Total	44		25 Unique genes

Table I -B *Receptor and Binding Protein Sequences*
Gene Abbreviation	Gene Frequency	UniGene Frequency	Putative and Known Functions
Il6st	1	0	cytokine receptor (including CNTF)
Lgals8	1	0	lectin receptor (galactose binding)
Leprot	1	0	leptin receptor (similar)
LOC383229	1	0	oxygen binding
Pmp22	1	75	myelin (peripheral)
Leprotl1	1	112	leptin receptor (similar)
D430039N05Rik	1	112	neuroreceptor (rhodopsin like)
Sorl1	2	150	LDL receptor (vacuolar sorting)
Fstl1	1	150	activin binding
Glra1	1	112�	glycine receptor widely distributed in CNS)
Gpr37	1	187�	neuroreceptor (rhodopsin like)
Adam23	1	262�	integrin ligand (possible neurogenesis)
Total	13		12 Unique Genes

Table I -C *Transport Protein Sequences*
Gene Abbreviation	Gene Frequency	UniGene Frequency	Putative and Known Functions
Slc3a2	1	0	amino acid transport
Pdzk1	1	0	Phosphate transport
Rhoq	1	0	glucose uptake (insulin stimulated)
Apol2	1	0	lipid transport
Hspa8	1	0	protein folding
Arl6ip1	1	0	protein transport (possible)
Dysf	1	0	vesicle fusion
Flot2	3	37	glucose uptake
1a2	2	37	sodium ion transport
Gnai3	1	37	potassium channel
Vamp1	4	75	vesicle fusion for transmission
Kcnk1	1	75	potassium channel
Vps25	1	112	endosomal sorting
Slc6a5	4	150�	glycine reuptake (myelin)
Trf	1	163	iron Transport (myelin)
Stxbp1	2	187�	vesicle docking Ach release
Nsf	1	300	vesicle fusion for transmission
Kcnj10	1	225�	potassium channel (myelin)
ApoD	1	412�	lipid transport
Total	29		19 Unique Genes

Table I -D *Regulatory Protein Sequences*
Gene Abbreviation	Gene Frequency	UniGene Frequency	Putative and Known Functions
Sel1h	1	[no data]	negatively modulates notch protein
Moap1	1	0	apoptosis
Rap1b	1	0	axon development
Stat5b	3	0	transcription regulation
Tspan7	1	0	neurite outgrowth putative
Rpl9	1	0	Ribosomal RNA binding
Dzip1	1	0	transcription regulation
Tcfcp2	1	37	Transcription regulation
Cstf2t	1	37	transcription regulation
Krba1	1	75�	transcription regulation
Hsp90b1	2	112	anti-apoptosis
Atxn2	1	112	apoptosis
Cdkn2d	1	112�	inhibits cyclin dependent kinase 2
Calcb	1	112�	neuropeptide activity
Abat	1	122	GABAase
Cnot2	1	150	transcription regulation
Ncor1	1	150	transcription regulation
Ncdn	1	150	bone resorption
Rpa1	1	187	DNA binding
Morf4l2	1	225	transcription regulation
Tbl1xr1	8	225�	transcription regulation
Ndrg1	1	262	axon survival
Rtn3	1	262	neural survival (co precipitates with Bace1)
s100b	4	262�	neural survival (located in astrocytes)
Hnrpf��	2	337�	Heterogeneous nuclear RNA binding
Total	39		25 Unique Genes

Table I -E *Enzyme Sequences*
Gene Abbreviation	Gene Frequency	UniGene Frequency	Putative and Known Functions
Lars2	26	0	Leucine aminoacyl transferase (translation)
Htatip	1	0	acetyltransferase histone (DNA repair)
Pgam2	1	0	glycolysis
Pld1	1	0	phospholipase
Lonrf2	1	0	protease
Stt3b	1	0	transferase
Rnf128	1	0	ubiquitin ligase
Plcb4	2	37	Phospholipase C
Ext1	1	37	glycosyl transferase
P4ha2	1	37	Prolyl 4-hydroxylase (collagen formation)
Pdhx	1	37	pyruvate dehydrogenase (PDH) complex component X
Pcmt1	2	75	methyltransferase carboxyl (protein)
Acsl5	1	75	acetyl CoA synthetase (lipid metabolism)
Pdha1	1	75	pyruvate decarboxylase (alpha subunit)
Pnpt1	1	75	RNA processing
Pgm2	2	112	glycolysis phosphoglucomutase activity
Mod1	1	112	glycolysis Dehydrogenase malate
Ndufs2	1	112	oxidoreductase NADH-ubiquinone
Aldoc	3	150	glycolysis aldolase 3, C isoform
Tnks2	1	150	transferase PARP activity
Ube3a	1	186	ubiquitin protein ligase
Scd2	1	487	Stearoyl-CoA desaturase (lipid metabolism)
Klk6	2	150�	protease
Fbxl2	1	300�	ubiquitin protein ligase activity
Bcat1	1	450�	transferase amino
Bace1	1	525�	protease APP (co precipitates with Rtn3)
Nat8l	2	825�	acetytransferase (drug metabolism) aka 1110038O08Rik
Total	59		27 Unique Genes

Table I -F *Hypothetical, Putative and Unknown Protein Sequences and Novel Expressed Sequence Tags*
Gene Abbreviation	Gene Frequency	UniGene Frequency	Putative and Known Functions
ES387449	1	[no data]	Chromosome 10 or Mitochondrial rnp (on the mitochondrial chromosome). Exhibits poly A therefore probably on chromosome 10 and not yet shown to be expressed i.e. no EST found in BLAST
ES387454	1	[no data]	Chromosome 8 (6402746-6403184) (new EST this study)
ES387457	1	[no data]	Chromosome 11 (20826626-20826095) (one additional EST #AW493294)
ES387458	1	[no data]	chromosome 1 (2145025-2145486) or Mitochondrial chromosome (11164-10586)� no ESTs found in BLAST
1110001P04Rik	1	0	Unclassifiable (putative)
B230209007	1	0	Unclassifiable (hypothetical protein)
D9Ertd402e	1	0	Unknown (hypothetical protein LOC382117)
Armcx1	1	0	Unknown (possible involvement in normal cell growth)
Ahi1	1	0	Unknown (WD40 domain, found in a number of eukaryotic �� proteins that cover a wide variety of functions including �� adaptor/regulatory modules in signal transduction, �� pre-mRNA processing and cytoskeleton assembly)
Brwd2	1	0	Unknown (WDR11 is a candidate gene for the frequently proposed tumor suppressor gene in 10q25-26 which is involved in tumorogenesis of glial and other tumors showing frequent alterations in the distal 10q region)
Dhcr24	1	0	Putative
Ubxd4	1	0	Unknown (hypothetical UBX domain containing protein)
2610204L23Rik	1	37	Hypothetical protein
4732416N19Rik	2	37	Unclassifiable (putative)
Dctn6	1	37	Unknown (putative mitochondria biogenesis)
Golph3	2	37	Unknown (putative trans Golgi protein)
Slain1	1	37	unknown
Rab11fip4	1	37	unknown
Tspan4	1	37	Unknown (complexes with integrins and possibly involved in cell development)
1200009O22Rik	1	75�	transferase (putative)
Cyhr1	1	75	Unknown (cystien and histidine rich protein 1)
E430025E21Rik	1	75	Unknown (hypothetical protein LOC223593)




Table I �F (cont.) *Hypothetical, Putative and Unknown Protein Sequences and Novel Expressed Sequence Tags*
Wdr20a	1	81	Unknown (WD40 domain, found in a number of eukaryotic �� proteins that cover a wide variety of functions including �� adaptor/regulatory modules in signal transduction, �� pre-mRNA processing and cytoskeleton assembly)
Tmem85	1	112	Unknown (hypothetical protein LOC68032)
Sh3bp5l	1	112	unknown
B230380D07Rik	1	122	Unknown (hypothetical protein LOC235461)
C030046I01Rik	1	187	Unknown (hypothetical protein, putative �� weakly similar to HYPOTHETICAL 4.9 KDA PROTEIN
Tnfrsf19l	1	262�	Unknown (putative)
Prpf19��	1	337�	Unknown (possible splicesome activity, possible ds DNA repair)
6330407J23Rik	1	337�	Unknown (hypothetical protein LOC67412)
6330527O06Rik	4	412�	Unknown (possible glycolsylation related �LAMP protein)
Total	36		32 Unique Genes

Listing of genes differentially expressed in spinal cord following subtraction with myelin cDNA arranged by UniGene frequency.�

1.� UniGene refers to the estimated number of transcripts per million listed for Mus musculus spinal cord as of March 31, 2007.

2.� (�) indicates genes that are among the three highest transcripts per million in spinal cord.�

TABLE II

GENES OF KNOWN GLIAL ORIGIN

DIFFERENTIALLY EXPRESSED

SPINAL CORD VERSUS WHITE MATTER

Table II
Genes of Known Glial Cell Origin Differentially Expressed in Spinal Cord Versus White Matter
Gene	Colonies	UniGene	Specificity
Slc6a5	4	150�	Solute carrier family 6 (neurotransmitter transporter, glycine), member 5
Gfap	3	112�	glial fibrillary acidic protein
Mpz	3	0	myelin protein zero
Plp1	2	712�	proteolopid protein
S100b	2	262�	S100 Calcium binding protein beta
Trf	1	163	transferrin
Mbp	1	2026�	myelin basic protein
Kcnj10	1	300�	Inwardly rectifying calcium channel subfamily J
Pmp22	1	75	peripheral myelin protein 22

Note 1:� Genes underlined and in italics are also seen in subtractions using visual cortex as driver and

spinal cord as tester (Lathia et al. 2006).

Note 2:� UniGene entries identified by � signify genes that are expressed either first, second or third

highest in spinal cord in Mus musculus.

TABLE III

GENE SEQUENCES

DIFFERENTIALLY EXPRESSED

WHITE MATTER VERSUS SPINAL CORD

Table III Gene Sequences Differentially Expressed in White Matter Versus Spinal Cord
	Colonies	Specificity
Ttr	11	Retinol binding (associated with amyloid disease)
Rasd2*	1	Cell communication membrane protein GTPase activity
Nos1ap	1	Studies of the related mouse and rat proteins have shown that this protein functions as an adapter protein linking nNOS to specific targets, such as Dexras1 and the synapsins.
D3Bwg0562e	1	peroxidase activity
Hpcal4	1	This encoded protein may be involved in the calcium-dependent regulation of rhodopsin phosphorylation. The transcript of this gene has multiple polyadenylation sites
Prkcb1	1	protein kinase
Txndc11	1	thioredoxin domain like 1
Zcchc12	1	Nucleotide binding zinc finger
Zmynd11	1	Nucleotide binding zinc finger
Atp2b1	1	Calcium transport membrane protein
Kcns2	1	Potassium channel
Itpr2	1	Inositol
M6prbp1	1	Protein phosphate 1
Psd2	1	Pleckstrin and Sec7 domain containing 2
Arpp-21	1	Cyclic AMP-regulated phosphoprotein, 21 kD
Dlgh4	1	scaffolding protein that binds and clusters N-methyl-D-aspartate receptors at neuronal synapses
Novel ESTs
Sub24-1rv2F7	1
Sub24-1rv2D8	1

Note. Listing of genes represented in order of number of colonies observed.�� Novel expressed sequence

tags have been submitted to the NCBI.

FIGURE 1

REPRESENTATIVE DOT BLOT ANALYSIS

Fig. 1 Representative dot blot analyses of four identical membranes spotted with 1uL of PCR product from individual bacterial colonies transformed with cDNA from forward suppression subtractive hybridization (SSH).� Each membrane was subsequently hybridized with the following probes:

Forward Subtracted signifies cDNA derived from spinal cord as tester and white matter as driver

Forward Unsubtracted signifies cDNA derived from spinal cord before subtraction against white matter.

Reverse Subtracted signifies cDNA derived from white matter as tester and spinal cord as driver.

Reverse Unsubtracted signifies cDNA derived from white matter before subtraction against spinal cord.

Clones are considered to be representative of cDNA sequences differentially expressed only if the density of labeling in the forward subtracted condition is at least twice that of any other condition.

FIGURE 2

REPRESENTATIVE NORTHERN ANALYSIS

Fig. 2 Northern blot analysis of total RNA from 90-110-day-old C57Bl6J male mice hybridized with radiolabeled probes.� The genes were selected from three categories according to their level of expression as recorded in the UniGene database.

Genes with no documented expression in spinal cord:� Secreted phosphoprotein 1-osteopontin (Spp1); Myelin protein zero (Mpz); RIKEN cDNA 111000P04 gene (111000P04Rik).
Genes that are moderately expressed in spinal cord: RIKEN cDNA 4732416N19 (4732416N19Rik); Peripheral myelin protein 22 (Pmp22); Flotillin 2 (Flot2); ATPase, Na+/K+ transporting; alpha 2 polypeptide (Atp1a2).
Genes highly expressed in spinal cord: Glial Fibrillary Acidic Protein (Gfap); Kallikrein 6 (Klk6); Neurofilament heavy chain (Nefh); Neurofilament medium chain (Nefm).

The samples were normalized with β-actin as represented in the figure.� Lanes in the figure are as follows:� SC represents spinal cord; WM represents white matter.� Spinal cord is the tester and white matter is the driver in the study.