Báo cáo khoa học: Lack of stabilized microtubules as a result of the absence of major maps in CAD cells does not preclude neurite formation pot

We have observed that the neurite microtubules of Cath.a-differentiated CAD cells, a mouse brain derived cell, are highly dynamic structures, and so we analyzed several aspects of the cy

of major maps in CAD cells does not preclude neurite

formation

C Gasto´n Bisig1, Marı´a E Chesta1, Guillermo G Zampar1, Silvia A Purro1, Vero´nica S Santander2 and Carlos A Arce1

1 Centro de Investigaciones en Quı´mica Biolo´gica de Co´rdoba (CIQUIBIC), UNC-CONICET, Departamento de Quı´mica Biolo´gica, Facultad de Ciencias Quı´micas, Universidad Nacional de Co´rdoba, Argentina

2 Departamento de Biologı´a Molecular, Facultad de Ciencias Exactas, Fı´sico-Quı´micas y Naturales, Universidad Nacional de Rı´o Cuarto, Argentina

Introduction

Correct functioning of the nervous system requires the

proper development of neuronal circuits and the

estab-lishment of synapses Although neurons from different regions of the nervous system acquire diverse

morpho-Keywords

CAD cells; microtubule-associated proteins;

microtubule dynamics; microtubules;

neurites

Correspondence

C A Arce, Centro de Investigaciones en

Quı´mica Biolo´gica de Co´rdoba (CIQUIBIC),

UNC-CONICET, Departamento de Quı´mica

Biolo´gica, Facultad de Ciencias Quı´micas,

Universidad Nacional de Co´rdoba,

5000-Co´rdoba, Argentina

Fax: +54 0351 4334074

Tel: +54 0351 000000

E-mail: caecra@dqb.fcq.unc.edu.ar

(Received 19 May 2009, revised

28 September 2009, accepted 2 October

2009)

doi:10.1111/j.1742-4658.2009.07422.x

In many laboratories, the requirement of microtubule-associated proteins (MAPs) and the stabilization of microtubules for the elongation of neurites has been intensively investigated, with controversial results being obtained

We have observed that the neurite microtubules of Cath.a-differentiated (CAD) cells, a mouse brain derived cell, are highly dynamic structures, and

so we analyzed several aspects of the cytoskeleton to investigate the molecu-lar causes of this phenomenon Microtubules and microfilaments were pres-ent in proportions similar to those found in brain tissue and were distributed similarly to those in normal neurons in culture Neurofilaments were also present Analysis of tubulin isospecies originating from post-trans-lational modifications revealed an increased amount of tyrosinated tubulin,

a diminished amount of the detyrosinated form and a lack of the Delta2 form This tyrosination pattern is in agreement with highly dynamic micro-tubules Using western blot analyses with specific antibodies, we found that CAD cells do not express several MAPs such as MAP1b, MAP2, Tau, dou-blecortin, and stable-tubule-only-peptide The presence of the genes corre-sponding to these MAPs was verified The absence of the correcorre-sponding mRNAs confirmed the lack of expression of these proteins The exception was Tau, whose mRNA was present Among the several MAPs investigated, LIS1 was the only one to be expressed in CAD cells In addition, we determined that neurites of CAD cells form and elongate at the same rate

as processes in a primary culture of hippocampal neurons Treatment with nocodazol precluded the formation of neurites, and induced the retraction

of previously formed neurites We conclude that the formation and elon-gation of neurites, at least in CAD cells, are dependent on microtubule integrity but not on their stabilization or the presence of MAPs

Abbreviations

CAD, Cath.a-differentiated; dCAD, diffentiated CAD; MAP, microtubule-associated protein; STOP, stable-tubule-only-peptide; TSA,

trichostatin A.

logies and abilities, there are certain basic features

common to all neurons (e.g the initiation and

elonga-tion of membrane protrusions for neurite formaelonga-tion,

and their stabilization and differentiation into

dendrites and axons) Other processes, such as the

organization of the internal cytoskeleton, migration,

guidance, and selective synaptogenesis, differ

depend-ing on the type of neuron [1,2]

The cytoskeleton is a critical structure for the

elon-gation of neurites and the maintenance of neuronal

architecture [3], and the stabilization of microtubules is

considered to be essential for neurons to maintain their

asymmetry and to transport materials required for

neurite elongation [4,5] The mechanism by which

neu-rons regulate microtubule assembly, stability, and

interactions with other cell structures is considered to

depend on the presence of microtubule-associated

pro-teins (MAPs), among which the most prominent are

MAP1b, MAP2, Tau, and stable-tubule-only-peptide

(STOP) [6–9] There are also other MAPs that have

been studied to a lesser extent (spectraplakins,

adeno-matous polyposis coli, doublecortin, dishevelled), as

well as other proteins binding to the plus end of

micro-tubules that could be involved in this process [10–12]

Transfection analyses have shown that Tau and

MAP2 induce the elongation of processes of

non-neu-ronal cells [13,14] Suppressed expression of MAP1b,

MAP2, and Tau using antisense and siRNA

technol-ogy in several studies [15–17] caused a reduction of

neurite outgrowth The microinjection of anti-Tau

antibodies into cultured neurons did not inhibit axonal

extension [18] Tau knockout mice showed a decreased

number of microtubules in small-diameter axons, but

extended axons were indistinguishable from those of

wild-type controls [19] MAP1b deficient mice show an

abnormal brain architecture, whereas, in MAP2

defi-cient mice, the cytoarchitecture was normal, suggesting

an overlapping function of MAP2 with MAP1b [20]

Lack of functional alteration in cases when only one

gene for MAP was silenced was generally attributed to

other proteins that provide additional redundancy with

MAP functions [21–25] The conflicting conclusions

made in different studies may be related to the use of

different technologies or different cell or tissue systems,

or the presence of MAPs with redundant functions In

any case, a requirement of MAPs and stabilized

micro-tubules for neurite formation has not yet been clearly

demonstrated In the present study, we characterized

cytoskeleton and neurite formation in

Cath.a-differen-tiated (CAD) cells, adding new information regarding

this particular subject

CAD originated as a subclone of the

cathecolamin-ergic cell line CATH.a, which was derived from a

neuronal brain tumor in a transgenic mouse expressing SV40 large T antigen under the control of the tyrosine hydroxylase promoter [26] CAD cells proliferate with

a rounded or polygonal shape in the presence of serum When serum is removed, they stop proliferating and differentiate, acquiring a neuron-like morphology, and, when serum is re-added, a rapid shortening of neurites is observed, such that most cells present a rounded morphology within approximately 40 min [27] Studies from several laboratories have shown that these cells contain synaptic vesicle proteins and express neuron-specific proteins such as b-tubulin III, GAP-43, SNAP-25, synaptostagmin, and other neuropeptides [27,28] The intracellular traffic powered by kinesins and dynein in these cells functions similarly to other neuronal systems [29–31] After differentiation, cell processes contain numerous varicosities similar to those of neurons [27,32] Single-cell electrophysiologi-cal studies have demonstrated that CAD cells can be induced to fire action potentials, and that voltage-dependent sodium and potassium currents can be elicited [33]

The rapid retraction of neurites after the addition of serum led us to consider the possibility that the cyto-skeleton of CAD cells should have peculiar properties Thus, in the present study, we investigated the main constituents of this structure and found that neurites have highly dynamic microtubules and lack stabilized microtubules because major MAPs are not expressed

in these cells However, neurites elongate at the same rate as those of normal neurons in culture

Results

Cytoskeletal proteins

As noted in the Introduction, the stabilization of microtubules is recognized as an essential process dur-ing the elongation of neurites, presumably to assure cell asymetry and the transport of materials to the growth cone Consistently, this process of stabilization has been described in several types of neurons in cul-ture [10] The rapid shortening of CAD cell neurites after the addition of serum led us to presume that there are alterations in the cytoskeleton Accordingly,

we investigated the presence, amount, and distribution

of the main components of this structure Immunofluo-rescence using specific antibodies revealed that actin microfilaments (Fig 1A) are present in CAD cells dis-playing the typical localization, positioned along the shaft and in the apical region of the growth cone pre-ceding the microtubules (Fig 1A) The three major cytoplasmic growth cone domains [i.e central (C),

transition (T) and peripheral (P) zones] can be clearly

distinguished (Fig 1A) The localization of these

pro-teins is the same as that previously reported in

cultured hyppocampal cells [10] Western blot and

subsequent determination of optical density confirmed

the presence of actin in CAD cells (Fig 1B, lanes 1

and 2) in an actin⁄ tubulin proportion slightly higher

than that determined in mouse brain [0.16 ± 0.05 and

0.12 ± 0.03 for diffentiated CAD (dCAD) cells and

brain, respectively, n = 3] We found no significant

difference in the amount of the neurofilament 100 kDa

constituent in relation to tubulin in CAD cells

com-pared to mouse brain (Fig 1B, lanes 3 and 4)

Acetylation and tyrosination states of tubulin The tubulin molecule is subject to a variety of post-translational modifications [14,34] One of them com-prises the reversible acetylation of its a-chain at the e-amino group of Lys40 [35] Although its physiological role is unclear, we have previously presented evidence demonstrating that acetylation is necessary for tubulin

to interact with Na,K-ATPase [36] In living cells, acet-ylated and deacetacet-ylated tubulin coexist in variable pro-portions depending on the cell type [37] Microtubules containing a high degree of acetylated tubulin were found to be more stable [35] In addition, microtubules

C

Fig 1 Tubulin, actin, and neurofilament protein expression in CAD cells (A) CAD cells differentiated for 5 days were stained for double immunofluorescence using rhoda-mine-conjugated phalloidin to detect actin microfilaments (Actin) and anti-total tubulin (Tubulin) to detect microtubules The merged image shows that actin microfila-ments invades the growth cone, whereas microtubules remain behind The central (C), transition (T) and peripheral (P) zones are also indicated Scale bar = 5 lm (B) CAD cells (80% confluence) were differentiated for 5 days, collected, and dissolved in Laemmli’s sample buffer for immunoblot in parallel with samples of mouse brain tissue Blots were stained simultaneously with anti-tubulin (DM1A) and anti-actin (lanes 1 and 2) Other samples were stained with anti-neurofilament protein (lanes 3 and 4) The volume of each sample was adjusted to load a similar amount of tubulin (C) CAD cells differentiated for 5 days were treated with 5 l M TSA for 0, 3, and 6 h, and imme-diately processed for immunofluorescence with anti-acetylated tubulin (clone 6-11B-1) (D) CAD cells differentiated for 5 days were treated with TSA for the indicated times and immunoblotted with anti-acetylated- and anti-total-tubulin (E) CAD cells differentiated for 5 days were treated for 12 h with 10 l M Taxol A control without Taxol was also run Cells were collected and processed for wes-tern blotting using antibodies against acety-lated and total tubulin The lane labeled +Taxol was overloaded to highlight the absence of acetylated tubulin.

were shown to be the preferred substrate for the

acety-lating enzyme [35] We found that the acetylated form

of tubulin was essentially absent in CAD cells (Fig 1C,

t= 0; Fig 1D, lane 0) This could be a result of the

predominance of highly dynamic microtubules versus

stable microtubules, to the predominance of

tubulin-deacetylase activity (histone tubulin-deacetylase 6) versus

tubu-lin-acetyltransferase activity, or to absence or inhibition

of the latter enzyme Treatment of cells with the

non-specific deacetylase inhibitor trichostatin A (TSA)

resulted in the appearance of a significant amount of

acetylated tubulin (Fig 1C, 3 and 6 h; Fig 1D),

indi-cating that both acetylase and deacetylase were present

in CAD cells Treatment of cells with Taxol induces an

increment in acetylated microtubules because the acetyl

transferase acts preferentially on these structures [35]

Stabilization of microtubules by treating CAD cells

with 10 lm Taxol did not cause increase of acetylated

tubulin (Fig 1E), indicating that the acetylation state

of tubulin depends mainly on the relative activities of

the acetylating and deacetylating enzymes rather than

on microtubule dynamics

Tyrosination⁄ detyrosination at the COOH-terminus

of a-tubulin is another post-translational modification

that has been extensively studied, although its

physio-logical role also remains unclear [38–40] As a result of

this cyclic modification, different isotypes of tubulin

exist: tyrosinated (Tyr-tubulin), detyrosinated

(Glu-tubulin), and Delta2 (a-tubulin lacking the two

COOH-terminal amino acids) Glu-tubulin and

Delta2-tubulin have been used as markers of stable

micro-tubules [41] Immunofluorescence images of CAD cells

using an antibody against total tubulin (which does

not discriminate different states of tubulin

tyrosina-tion) showed a bright, typical microtubule network in

the cell body and neurites (Fig 2A) A similar pattern

was observed using an antibody specific to tyrosinated

tubulin Antibody against Glu-tubulin revealed scarce,

curly microtubules, whereas antibody against

Delta2-tubulin revealed no microtubules These results were

confirmed by immunoblots using the same antibodies

(Fig 2B) Mouse brain tissue was used as a positive

control The reduced amount of Glu-tubulin in CAD

cells was not a result of a lack of (or inhibition of) the

putative detyrosinating enzyme (tubulin

carboxypepti-dase) because a significant increase of Glu-tubulin was

observed in differentiated and nondifferentiated cells

treated with Taxol (Fig 2C)

Microtubule dynamics

The rapid shortening of neurites found in CAD cells,

along with the absence of markers of stable

micro-tubules (Glu-tubulin, and Delta2-tubulin), led us to consider the possibility that microtubules are highly dynamic structures in these cells By measuring the rate of microtubule depolymerization after nocodazole treatment [4,5], microtubule dynamics in CAD cells was compared with that of other cell types Micro-tubules of CAD cells were as dynamic as those of Chinese hamster ovary and PC12 cells in active prolif-eration The time required for 50% depolymerization was 1–2 min (Fig 3A, B, empty circles) On the other hand, the depolymerization curve for 7-day-old chicken embryo brain cells showed a two-phase behav-ior, suggesting the presence of two microtubule popu-lations: one with a half-life of 1–2 min and the other being more stable (Fig 3A, B, solid triangle) As a negative control of microtubule disassembly by noco-dazole treatment, CAD cells were pre-treated with sodium azide, which stabilizes microtubules by deplet-ing cells of ATP [42] Under these conditions, microtu-bules were not disassembled by nocodazole treatment (Fig 3A, bottom; Fig 3B, solid circles)

Several MAPs are not expressed in CAD cells From a mechanistic point of view, there is a general consensus that MAPs are the proteins responsible of microtubule stabilization [10,22,43,44] Thus, we investigated whether the occurrence of highly dynamic microtubules in CAD cells is the result of some alter-ation in one or more MAPs The presence of neuro-nal structural MAPs (i.e MAP1b, MAP2, Tau, and STOP) was investigated in 10-day-differentiated CAD cells by immunoblotting using appropriate antibodies

In the case of Tau, immunoblots were revealed with antibodies that recognize dephosphorylated and phos-phorylated epitopes and a nonphosphorylable region

of the protein (Tau-1, Tau-2 and 134d) For compari-son, soluble fractions from 30-day-old mouse brain were simultaneously run All the MAPs investigated were present in brain samples, but not in samples from CAD cells (Fig 4) Brain and CAD samples run in each lane contained similar amounts of a-tubulin (Fig 4, lower panels) The experiment was repeated, running overloaded samples of CAD cells and using a more sensitive chemiluminescent method (Femtomolar detection system), with similar results being obtained (i.e no band was observed in lanes corresponding to CAD cells) This is exemplified by

an overloaded dCAD cell sample being revealed with 134d antibody (Fig 4, lane dCAD⁄ Overload) More-over, treatment of nitrocellulose membrane with alka-line phosphatase prior to incubation with anti-Tau-1, aiming to increase the epitopes that can be

recognized, produced a significant increase in Tau

bands in the Br lane, but no band appeared in the

dCAD lane (Fig S1)

To concentrate MAPs eventually diluted in the cell

extract, we performed immunoprecipitation with

Sepharose beads linked to antibodies specific to each

MAP As a control, mouse brain samples were also

analyzed in parallel The amount of the brain soluble

fraction and dCAD cell extract used in these

experi-ments as input material was 30-fold higher than those

loaded on each lane shown in Fig 4 For each MAP,

most of the protein in brain samples was found in the

pellet, whereas, in dCAD cells samples, no MAP band was observed (not shown)

Gene and mRNA analyses of MAP1b, MAP2, Tau, and STOP

The finding that apparently normal neurites are formed even when CAD cells lack MAP1b, MAP2, Tau, and STOP proteins was surprising This led us to investigate the presence of their respective genes and messenger RNAs, using a PCR technique with specifically designed primers (Table 1) In every case, the PCR products

A

C

B

Fig 2 Tyrosination state of tubulin (A) Cells differentiated for 5 days (dCAD) and nondifferentiated cells (CAD) were visual-ized by immunofluorescence using antibod-ies specific to a-tubulin (total tubulin), Tyr-, Glu-, and Delta2-tubulin The inset shows embryonic chicken brain cells differentiated for 6 days in culture and revealed with anti-body to Delta2-tubulin Scale bar = 10 lm (B) Cells obtained as in (A) were subjected

to western blotting and stained with the same antibodies as in (C) For staining with each antibody, identical volumes of CAD cell samples were run As positive controls, samples of mouse brain (Br) were included (C) Nondifferentiated CAD cells and cells differentiated for 7 days were treated (+) or not ( )) with 10 l M Taxol for 12 h and then subjected to western blotting using antibody to detyrosinated tubulin (Glu-tubulin) All lanes were loaded with samples containing the same amount of total tubulin.

A B

Fig 3 Nocodazole sensitivity of microtubules of CAD and different cell types CAD cells differentiated for 7 days, PC12 cells (80% conflu-ence), Chinese hamster ovary cells (80% confluconflu-ence), a primary culture of 7-day-old chicken brain cells, and differentiated CAD cells treated with 20 m M sodium azide (in culture medium without glucose) for 1 h, were incubated in the presence of 10 l M nocodazole for the indicated times and immediately processed to isolate the cytoskeletal fraction, which remained bound to the plastic dish (see Experimental proce-dures) The cytoskeletal fraction remaining after nocodazole treatment was processed for western blotting and stained with antibodies to total tubulin (DM1A) and actin (as a loading control) (A) Immunoblots from a typical experiment (B) Optical density values for total tubulin corresponding to bands from three independent experiments (mean ± SD) For each type of cell, the attenuance of the tubulin band at time zero of nocodazole treatment is considered to be 100%.

Fig 4 Analysis of microtubule-associated proteins in differentiated CAD cells CAD cells were grown on 10 cm dishes (to 80% confluence) and differentiated over 10 days in fetal bovine serum-free culture medium Cells were collected, dissolved in a small volume of Laemli’s sample buffer, and subjected to SDS ⁄ PAGE (6% acrylamide for MAP1B, MAP2 and STOP; and 10% for Tau) and immunoblotting using anti-bodies to various MAPs as indicated As positive controls, samples of supernatant fractions from mouse brain homogenates centrifuged at

100 000 g were processed in parallel (Br) and revealed with antibodies to each of the MAPs For brain and CAD cells, the volume loaded in each lane was adjusted to contain equivalent amounts of total tubulin, as revealed with DM1A antibody (bottom panel), except for the lane

on the right, which was revealed with 134d (dCAD⁄ Overload) in which a triple amount of total tubulin was loaded The positions of mole-cular mass markers are indicated.

obtained represent approximately 30% of the complete

genes At least these portions of the genes corresponding

to each of the MAPs were present in CAD cells

(Fig 5A) We consider it most likely that the complete

sequences of the respective genes are present in the cell

genome because it would be an extreme coincidence that

the rest of each gene had been missed Analysis of the

respective mRNAs by RT-PCR, using the same primers,

indicated that the mRNAs of MAP1b, MAP2, and

STOP were absent in nondifferentiated cells, whereas, in

differentiated cells, a weak band was observed for

MAP1b and STOP On the other hand, the quantity of

RT-PCR product corresponding to Tau in both

differ-entiated and nondifferdiffer-entiated CAD cells was similar to

that in brain tissue (Fig 5B)

LIS1 but not doublecortin is expressed in CAD

cells

Other proteins, such as a-Lis 1 and doublecortin, have

been shown to interact, directly or indirectly, with

microtubules and to stabilize them in vitro [11,12,45]

Investigations on the biochemical basis of lissencephaly,

a human neurological disease characterized by an

abnormal layering of brain cortex, led to the discovery

of these two proteins, which are lacking or mutated in

patients [46] Although they are not major MAPs of

neurons (based on their quantity in total brain), we

investigated their presence in CAD cells Immunoblots

using the corresponding antibodies revealed the

presence of a-Lis 1 and the absence of doublecortin in

these cells (Fig 6A) Similarly, RT-PCR using

specifi-cally designed primers (Table 1) revealed the absence

of mRNA corresponding to doublecortin and the

presence of a-Lis 1 mRNA (Fig 6B)

Neurite formation in CAD cells CAD cells were grown under differentiating conditions

as described previously [27], microphotographs were taken on various days, and neurite length was mea-sured On day 0, cells were rounded, with only minor membrane protrusions Numerous processes subse-quently appeared, and grew rapidly to form a dense meshwork (Fig 7A, 15 days) Varicosities, similar to those of neurons in primary culture, were observed in all of the processes (not shown) On day 15, cells were changed to culture medium containing 10% fetal bovine serum, and photographed 24 h later (Fig 7A, +FBS 24 h) As reported previously [27], fetal bovine serum treatment induces the retraction of processes, and cells assume a rounded or polygonal form with scarce, short processes and resume proliferation (not shown) Neurite length was quantified as a function of days in culture under differentiating conditions (Fig 7B) During days 1–8, neurites elongated at an average rate of approximately 40 lm per day This rate is very similar to that of axons in central nervous system cells in culture [47,48] For statistical measure-ment of neurite retraction, at day 7 under differentiat-ing conditions, cells were changed to culture medium containing 10% fetal bovine serum, and cultured for

an additional 24 h Neurite length determination demonstrated that the processes retracted almost completely (Fig 7B, open square)

The peculiar properties of the CAD cell cytoskeleton compelled us to investigate to what extent neurite for-mation is a microtubule-dependent process We found that treatment of nondifferentiated cells with nocodaz-ole precluded neurite outgrowth, and a similar treat-ment after differentiation led to the retraction of

Table 1 PCR primer sequences used for screening expression of different MAPs genes by CAD cells.

MAP1b-for

MAP1b-rev

GAGCTGGAGCCAGTTGAGAAGCAGGG GTTGGTCTCGTCGCTCATCACATCACGAGG

82898–82923 83581–83552

NC_000076 Idem MAP2-for

MAP2-rev

GCTTGAAGGCGCTGGATCTGCGACAATAG GACTGGGCTTTCATCAGCGACAGGTGGC

91489–91517 92431–92404

NC_000067 Idem Tau-for

Tau-rev

GTGAACCACCAAAATCGGAGAACGAAGC CAGGTTCTCAGTAGAGCCAATCTTCGACCTGAC

78772–78800 79013–78981

NC_000077 Idem STOP-for

STOP-rev

AGAGTCGGATGCAGTTGCCCGGGCAACA GGCTCCTCCAGCACCCTCCGGGTCCCG

210–237 657–631

NC_000073 Idem Doublecortin-for

Doublecortin-rev

CCCCAAACTTGTGACCATCATTC GGAGAAATCATCTTGAGCATAGCG

705–728 967–943

NM_010025 Idem LIS1-for

LIS1-rev

CGAACTCTCAAGGGC ATGCATCAGAACCATGCACG

1288–1303 1427–1407

NM_95116 Idem Tubulin a6-for

Tubulin a6-rev

AGCCCTACAATTCCATCCTCACC GCTGAAGGAGACGATGAGGGTGA

6854–6876 7646–7624

NC_000081 Idem

neurites (results not shown), indicating that

micro-tubule integrity is necessary for both elongation and

sustaining neurites

Discussion

Our understanding of neurogenesis, neuronal plasticity,

and the establishment of correct synapses and circuits

in the central and peripheral nervous systems has

advanced greatly over the past decade The most

stud-ied MAPs (i.e MAP1b, MAP2, Tau, and STOP) have

been shown to promote the polymerization and

stabil-ization of microtubules, and therefore these proteins

and microtubules are involved in the elongation of

neural processes (i.e the establishment of neuronal

polarity) [10,22,43,44]

We found that MAP1b, MAP2, Tau, STOP, and

doublecortin are not expressed in CAD cells (Fig 4)

This was observed by an immunoblot using specific

antibodies against each MAP Complementary

experi-ments [immunoprecipitation, overloaded gels, highly

sensitive chemiluminescent method (Femtomolar

detection system) and the use of different antibodies

against Tau] confirmed the absence of these proteins

Molecular biology techniques showed the presence of

the genes corresponding to each MAP and the absence

of their mRNAs (with the exception of that of Tau)

(Fig 5) mRNA corresponding to Tau was detected in

CAD cells in amounts similar to that in brain tissue

(Fig 5), suggesting that Tau expression is inhibited at the translational level, whereas other MAPs are down-regulated at the transcriptional level

A study showing the expression of MAP1b in CAD cells using a polyclonal antibody was recently published [32] However, when we tested the same antibody (a gift from I Fisher, Drexel University, Philadelphia, PA, USA) on either mouse brain or CAD cells samples, we obtained a complex and con-fusing pattern of bands (not shown) Thus, we were unable to draw any conclusions regarding this anti-body This observation, in addition to the absence of any band on the immunoblot stained with a com-mercial anti-MAP1b (Fig 4) and the strong evidence about the absence of MAP1b mRNA (Fig 5), leads

us to conclude that MAP1b is not expressed in CAD cells Even if this protein were expressed at a very low level, as suggested by the trace amount of MAP1b mRNA shown in Fig 5B, it is evident (from the results provided in Fig 3) that the amount of this MAP is insufficient to stabilize microtubules Tubulin, actin, neurofilament protein (Fig 1), LIS1 (Fig 6), and the other proteins tested (not shown) are present in CAD cells in normal amounts and with nor-mal cellular distribution, suggesting that these proteins are not involved in the mechanism that leads to the peculiar behaviour of CAD cells It is a remarkable coincidence that only those proteins having the ability

to associate directly with microtubules (structural

Fig 5 Analysis of genes and mRNAs corresponding to MAP1b, MAP2, Tau, and STOP in CAD cells (A) Genomic DNA from CAD cells dif-ferentiated for 10 days, and from mouse brain, was purified and subjected to PCR using primers specifically designed to detect each of the MAPs (see Experimental procedures and Table 1) Products were electrophoretically separated on agarose gels and stained with ethidium bromide For each MAP, single bands were obtained in each lane Standard molecular masses are shown on the right (B) Total RNA from mouse brain and 10 day-differentiated (dCAD) and nondifferentiated (CAD) cells were purified and subjected to RT-PCR with the same prim-ers used in (A) As a positive control of expression, primprim-ers designed to detect the presence of a-tubulin 6 mRNA (a protein of constitutive expression) were also used (Table 1).

MAPs) and stabilize them are absent in CAD cells A

possible explanation is that the expression of all these

MAPs is under a common regulatory mechanism

Alternatively, the expression of each MAP could be

sequential, so that the expression of each MAP would

depend on the regulation of the previous one in the

sequence

Dynamic and stable microtubules coexist in

neu-rons For example, Baas et al [49] reported a half-life

of 3.5 and 130 min for dynamic and stable subpopu-lations, respectively Proximal microtubules in axons are more stable than distal ones [50], suggesting that microtubules become stabilized as the process elon-gates On the basis of sensitivity to nocodazol treat-ment, microtubules in CAD cells were shown to be highly dynamic (half-life = 2 min) (Fig 3) Similarly, these microtubules contain a very low level of detyro-sinated tubulin and no Delta2 tubulin, which are markers of stable microtubules (Fig 1C, D) Further-more, the level of tyrosinated tubulin (a marker of dynamic microtubules) was high (Fig 1C, D) Taken together, these results clearly indicate that microtu-bules in CAD cells are highly dynamic structures This is consistent with the lack of microtubule-stabi-lizing MAPs in these cells

The hypothesis underlying most of the numerous experiments that have been performed to elucidate the physiological role of MAPs assumes that these proteins stabilize microtubules, and thus are therefore required for the extension of membrane protrusions such as axons and dendrites We found that apparently normal neurites in CAD cells elongate similarly to neurites in primary culture (Fig 7), even though the microtubules lack most MAPs (Figs 4 and 5), and are highly dynamic structures (Fig 3) With regard to neurite elongation, MAPs could theoretically be ‘substituted’

by other yet-undescribed proteins having redundant functions However, the finding in the present study that microtubules in CAD cells are highly dynamic indicates that no mechanism is operating to compen-sate for the absence of the microtubule-stabilizing function of MAPs

The results obtained in the present study are consis-tent with the idea that even though intact microtubules are necessary for neurite elongation, neither stabiliza-tion of these structures nor the presence of MAPs is required The only MAP that we found to be expressed in CAD cells is LIS1 (Fig 6) This protein belongs to a unique class of microtubule-binding pro-teins termed +TIPS (for plus-end tracking propro-teins) [51] and is a regulated adapter between CLIP-170 and cytoplasmic dynein In addition, LIS1 forming a com-plex with other proteins (e.g dynein⁄ dynactin and Clip170) was suggested to be necessary for the elonga-tion of the growth cone, cell migraelonga-tion, prevenelonga-tion of catastrophe events, docking of the growing microtu-bule to specific cortical sites, tethering microtumicrotu-bules to the cell cortex, etc [45,52,53] In this scenario, we can imagine that the +TIPs complex is responsible for the elongation of the neural processes without the need for microtubule stabilization or the expression of struc-tural MAPs In normal neurons, MAPs may regulate

B

A

Fig 6 LIS1 but not doublecortin is expressed in CAD cells (A)

Dif-ferentiated (dCAD) and nondifDif-ferentiated (CAD) cells were

sub-jected to SDS ⁄ PAGE and immunoblot with antibodies to

doublecortin (A, left) and to LIS1 (A, right) As positive controls,

samples of cytosolic fractions from adult or newborn mouse brain

(for LIS1 or doublecortin, respectively) were included (Br) For

com-parison, total tubulin (as revealed with the monoclonal DM1A

anti-body) contained in each sample was also determined (A, bottom

panel) (B) Total RNA from mouse brain (Br) and 10 day-dCAD cells

were purified and subjected to RT-PCR with primers specifically

designed to detect doublecortin or a-Lis 1 (Table 1) After 46 cycles

of PCR, samples were loaded in an agarose gel, and stained with

ethidium bromide.

microtubule dynamics not for the purpose of initiating

or sustaining neurite elongation, but to modulate other

more subtle functions (e.g spatial organization of

microtubules, interaction with other structures, growth

cone guidance, synaptogenesis, etc.) Because five

major MAPs are absent in CAD cells, these cells

provide a useful model for studying the roles of

other cytoskeletal proteins in neurite formation at the

molecular level

Experimental procedures

Chemicals

Nocodazole, paclitaxel (Taxol), TSA,

rhodamine-conju-gated phalloidin, sodium butyrate, and culture media were

obtained from Sigma-Aldrich (St Louis, MO, USA) Fetal

bovine serum was obtained from Natocor (Co´rdoba,

Argentina)

Soluble mouse brain extract preparation

Brains from 15- to 30-day-old mice were homogenized in

1 vol (w⁄ v) of cold MEM buffer (100 mm Mes adjusted

with NaOH to pH 6.7, containing 1 mm EGTA and

1 mm MgCl2) The homogenate was centrifuged at

100 000 g for 1 h, and the supernatant fraction was

col-lected

Cell culture

Brain cells from 7-day-old chicken embryos were isolated

and cultured as described previously [54] Chinese hamster

ovary and PC12 cells were grown in DMEM containing

10% fetal bovine serum (fetal bovine serum) at 37C in an

air⁄ CO2 (19 : 1) incubator CAD cells were grown on

35 mm dishes in DMEM⁄ F12 (50 : 50, v ⁄ v) with 10% fetal

bovine serum and 2 mm glutamine The differentiation of

these cells was accomplished by replacing the medium with

the same medium lacking fetal bovine serum Under these

conditions, neurites longer than five soma diameters are

visualized after 24–48 h In all experiments, the

differen-tiation status of cells was confirmed by microscopic

exami-nation

Antibodies

Rabbit polyclonal antibodies specific to Glu-tubulin

(anti-Glu) and to Delta2-tubulin were prepared in our laboratory

as described previously [55] Mouse monoclonal antibodies

against Tyr-tubulin (Tub 1A2, 1 : 1000), total a-tubulin

(DM1A, 1 : 1000), b-actin (Clone AC-15; 1 : 500),

acety-lated tubulin (6-11B-1, 1 : 1000), peroxidase-conjugated

rabbit anti-(mouse IgG) (1 : 800), rhodamine-conjugated

goat anti-(rabbit IgG) (1 : 600) and fluorescein-conjugated goat anti-(mouse IgG) (1 : 600) were obtained from Sigma-Aldrich Mouse monoclonal antibody mainly specific

B

A

8 days

3 days

Fig 7 Elongation and retraction of neurites in CAD cells CAD cells were grown under proliferating conditions on coverslips, up to approximately 40% confluence, and transferred to culture medium without fetal bovine serum (FBS) (A) Images were taken from 0–15 days of differentiation At day 15, fetal bovine serum was added (10% final concentration), and cells were photographed 24 h later Scale bar = 100 lm (B) At the indicated days of culture, five different areas from three different plates were analyzed to mea-sure the length of the processes The sum of the lengths of all the measured processes was divided by the number of cells Cells with

no process were excluded from the analysis At day 7 under differ-entiating conditions, cells were changed to culture medium contain-ing 10% fetal bovine serum and, after 24 h, neurite length was measured as described above (open square) Values are the mean ± SD of three independent experiments.