Identification and characterization of novel genes modulated in the thyroid of dogs treated with methimazole and propylthiouracil.

Induction of cell proliferation by mitogen or growth factor stimulation leads to the specific stimulation or repression of a large number of genes. To better understand differentiated epithelial cell growth regulation, we have initiated a study to identify genes which are regulated by the thyrotropin-dependent mitogenic pathway in dog thyroid cells. A thyroid cDNA library was prepared from a methimazole and propylthiouracil-treated dog and differentially screened with probes derived from control or stimulated thyroids. Among 19 clones isolated, 6 encode known proteins (inwardly rectifying potassium channel, nucleosome assembly protein, ribosomal protein L7, elongation factor 1α, non-muscle myosin light chain, and heat shock protein 90β). The 13 others correspond to proteins whose function is unknown. Among them, 5 correspond to mRNAs whose expression was modulated by mitogenic stimulation of thyrocytes in primary culture. A preliminary characterization of two of these cDNAs is reported: clone 5, which might represent a novel, atypical protein kinase, and clone 3, which contains ankyrin-like repeats, suggesting that it might interact with other proteins.

Induction of cell proliferation by mitogen or growth factor stimulation leads to the specific stimulation or repression of a large number of genes. To better understand differentiated epithelial cell growth regulation, we have initiated a study to identify genes which are regulated by the thyrotropin-dependent mitogenic pathway in dog thyroid cells. A thyroid cDNA library was prepared from a methimazole and propylthiouraciltreated dog and differentially screened with probes derived from control or stimulated thyroids. Among 19 clones isolated, 6 encode known proteins (inwardly rectifying potassium channel, nucleosome assembly protein, ribosomal protein L7, elongation factor 1␣, nonmuscle myosin light chain, and heat shock protein 90␤). The 13 others correspond to proteins whose function is unknown. Among them, 5 correspond to mRNAs whose expression was modulated by mitogenic stimulation of thyrocytes in primary culture. A preliminary characterization of two of these cDNAs is reported: clone 5, which might represent a novel, atypical protein kinase, and clone 3, which contains ankyrin-like repeats, suggesting that it might interact with other proteins.
Induction of cell proliferation by mitogen or growth factor stimulation leads to the specific and sequential expression of a large number of genes (1)(2)(3)(4)(5)(6)(7)(8): immediate early genes, induced rapidly and independently of de novo protein synthesis, delayed early genes whose transcriptional activation requires new protein synthesis, and others ending with the late G 1 genes, which are mainly enzymes involved in DNA synthesis. Conversely, the expression of other genes is also sequentially repressed after mitogenic stimulation (9 -11). Many of these regulated genes are proto-oncogenes or anti-oncogenes. Most of the known ones belong to two major classes of mitogenic pathways: the growth factors receptors tyrosine protein kinases and the phorbol ester protein kinases C cascades. A third signaling pathway, the cyclic AMP-dependent cascade, considered for a long time to be a general inhibitor of proliferation, stimulates proliferation in several epithelial cell types and in yeast (reviewed in Ref. 12). In thyroid cells, this pathway is activated by thyrotropin (TSH), 1 which is the main physiological agent regulating the thyroid gland (13,14). TSH and cyclic AMP promote cell proliferation, function, and differentiation while the mitogenic pathways elicited by EGF and tumor promoting phorbol esters are associated with the loss of expression of the differentiation specific genes (15)(16)(17)(18)(19). Several steps of the latter cascades are missing in the cyclic AMP pathway (14). The molecular mechanisms of this pathway are still largely unknown. We have therefore initiated a study aimed at the identification of genes that are regulated by cyclic AMP in thyroid cells.
To identify new genes that might be regulated by the thyroid mitogenic pathways, we prepared a cDNA library from the thyroid of a dog treated in vivo with methimazole and propylthiouracil. The inhibition by these drugs of thyroid hormones synthesis leads to a relief of the inhibitory control of these hormones on the hypophysis, to TSH secretion, and chronic stimulation of the thyroid. By differential screening, we identified 19 clones of which 13 have not been described to date. Among those, five corresponded to mRNA whose expression showed modulation after different mitogenic stimulations of thyroid cells in primary culture.

EXPERIMENTAL PROCEDURES
Animal tissues used for this work are the same as those of a previous study conducted in our laboratory (19). The dogs used in this study were used for cardiological studies, and their thyroids were resected before these experiments. Briefly, dogs were treated by oral administration for 4 weeks of methimazole (MM) (2 ϫ 60 mg/day strumazol) and propylthiouracil (PTU) (2 ϫ 150 mg/day propylthiurit) in order to increase the circulating TSH level. Triiodothyronine and thyroxine concentrations in the serum were followed by radioimmunoassay to ensure that the treatment had been effective. On the day of the experiment, 1 h prior to thyroid resection, the dogs received 50 mg/kg bromodeoxyuridine by intravenous injection. Bromodeoxyuridine labeling analyses were performed as described (20). Tissues for PCNA immunohistochemistry and RNA preparation were obtained from MM/PTU treated and untreated dogs. The immunohistochemical procedure for PCNA immunohistochemistry was used as described previously (21). The method involves methanol fixation of the tissue (previously snap-frozen in this case), followed by embedding in paraffin.
Dog thyroid follicles were obtained as detailed previously (22). The follicles were seeded in 100-mm Petri dishes and cultured in a control medium consisting of Dulbecco's minimal essential medium (Life Tech-* This work was supported by the Ministère de la Politique Scientifique (PAI), Fonds National de la Recherche Scientifique, Fonds de la Recherche Scientifique Médicale, European Union Biomed and Human Capital programs, Opération Télévie, Association Contre le Cancer, Association Sportive Contre le Cancer, and the Fondation Hoguet. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ These authors contributed equally to this study. Fellows of the Fonds pour la formation à la Recherche dans l'Industrie et l'Agriculture (FRIA).
To purify the RNA, thyroid tissues were ground into a fine powder under liquid nitrogen and homogenized in 4 M guanidinium thiocyanate; thyroid cells were scraped from culture dishes in the same solution. Lysed cells were layered over a cushion of 5.7 M cesium chloride in an SW56 tube and centrifuged at 36,000 rpm at 15°C for 17 h. Total RNA was extracted as described (23). After treatment with 60 g/ml proteinase K in 0.3 M NaCl, 0.1% SDS and phenol/chloroform extractions, the RNA was precipitated and resuspended in water for spectrophotometric quantification. Polyadenylated RNA was purified by oligo(dT)-cellulose chromatography (Stratagene) or by the PolyATtract system (Promega) following the manufacturer's protocol. The quality of the poly(A) ϩ RNA prepared for the construction of the library and for the synthesis of the probes was confirmed by Northern blot analysis, using a thyroglobulin probe, detecting a 8.5-kb mRNA. For tissue distribution analysis, total RNA of various dog tissues was extracted with the TRIzol Reagent kit (Life Technologies, Inc., Paisley, UK) (24).
For the construction of a cDNA library in ZAP II, the poly(A) ϩ RNA prepared from the thyroid of a MM/PTU-treated dog was used for the construction of an oligo(dT)-primed, directional cDNA library with the ZAP-cDNA synthesis and cloning system of Stratagene, according to the manufacturer's protocol. The resultant library contained approximately 10 7 primary recombinants with 0.7% non-recombinant phages. The library was amplified once by the plate lysate method to obtain 2.4 ϫ 10 10 plaque-forming units/ml. To assess size distribution of inserts in the cDNA library, individual plaques were randomly selected and plate lysate stocks were generated by infection of XL-1 blue Escherichia coli cells. Phage DNA was extracted and restricted by EcoRI/XhoI, and this was then electrophoresed on 1% agarose gels to visualize insert sizes. The mean size of the randomly selected cloned inserts was 2.6 kb.
Poly(A) ϩ RNA from in vivo quiescent or stimulated thyroid tissues was used as template for the synthesis of single-stranded digoxigeninlabeled cDNA probes using hexamers and Moloney murine leukemia virus reverse transcriptase following the protocol suggested by the manufacturer (Dog Labeling kit, Boehringer Mannheim). For differential screening, the library was plated at low density (approximately 6000 plaque-forming units/13-cm diameter dish), and duplicate nylon membranes (Qiabrane, Qiagen) lifts were made from each dish. Filters were treated with an alkaline solution, followed by neutralization and nucleic acid was fixed onto the membranes by UV cross-linking (25). Differential hybridization screening of the library was performed according to the procedure of the manufacturer (Boehringer Mannheim). The first filter was hybridized with the probe derived from quiescent thyroids, and the second one with the probe derived from stimulated thyroids. Following 4 h of prehybridization in 5 ϫ SSC, 1% (w/v) blocking reagent for nucleic acid (Boehringer Mannheim), 0.1% Nlauroylsarcosine, 0.02% SDS, the two blots were incubated for 16 h at 68°C in the hybridization solution containing the digoxigenin-labeled cDNA probes (25 ng/ml). Stringency washes were performed for 2 ϫ 10 min at room temperature with 2 ϫ SSC, 0.1% SDS, followed by 2 ϫ 20 min at 65°C with 2 ϫ SSC, 0.1% SDS and 2 ϫ 20 min at 65°C with 0.1 ϫ SSC, 0.1% SDS. The immunological detection of the digoxigeninlabeled probes was performed by a secondary alkaline phosphataselabeled antibody complex in combination with a chemiluminescent substrate LumigenPPD. Filters were exposed to Hyperfilm MP (Amersham Corp.) for 1-4 h at room temperature. Clones that hybridized differentially with the two probes were replated at low density for a second screening. The remaining positives were picked out, plaque purified, and further analyzed by Northern blotting and sequencing.
After in vivo excision of Bluescript phagemid from Uni-ZAP XR vector using the manufacturer's protocol, the resulting plasmid DNA was amplified and purified by the alkaline lysis method (25). Various restriction enzymes were used to map inserts and subfragments were further cloned in both orientations into M13 mp18 and M13 mp19 single-stranded DNA vectors. Sequence determination was performed by the dideoxy chain termination method (26) using Taq DNA polymerase and M13 fluorescent primers. Reactions were run on the Applied Biosystems 370A automated DNA sequencer. Approximately 300 base pairs of the extremities or the internal sequences of clones were compared with non-redundant European Molecular Biology Laboratory and GenBank data bases, using the BLAST Network Service. Sequences were also translated and used to search the protein data bases (Protein Identification Resource, SWISSProt, and GenPept) using the BLASTX sequence analysis program.

Construction of a cDNA Library from a Thyroid of a MM/
PTU-treated Dog-To construct a TSH-stimulated thyroid cDNA library, a dog was treated with methimazole and propylthiouracil as described (see above). At the time of thyroid resection, the dog was biologically hypothyroid and its thyroid was thus chronically stimulated by TSH. Serum thyroxine and triiodothyronine levels were 0.5 g/dl and Ͻ30 ng/dl, respectively, versus 1.1 Ϯ 0.1 g/dl and 37 Ϯ 6 ng/dl in control dogs. Histological examination showed markedly a hyperplastic thyroid with almost total absence of colloid in the antithyroid drug-treated animal, and normal follicles containing colloid in the control thyroids. Bromodeoxyuridine labeling analysis demonstrated nuclear labeling in 8.1% of the thyroid cells of the treated dog, whereas the thyroid cells of control animals showed only 0.17 Ϯ 0.07% labeled nuclei (mean Ϯ S.E.).
Total RNA was extracted from the thyroid of the MM/PTUtreated dog and the poly(A) ϩ RNA fraction was isolated. The quality of the poly(A) ϩ RNA was confirmed by Northern blot analysis, using a thyroglobulin probe, which detected an intact 8.5-kb mRNA (data not shown). This poly(A) ϩ RNA was used as template for the construction of an oligo(dT)-primed cDNA library in the ZAP II phage vector.
Identification of Modulated Genes by Differential Screening-The screening of 250,000 plaques, with cDNA probes from control and stimulated thyroids, yielded 19 plaques which, after a second screening, showed consistent differential hybridization. Among these, 11 showed overexpression in the treated thyroid, and 8 showed underexpression. All cDNA inserts were partially sequenced and the sequences were compared to the GenBank and EMBL data bases, using the BLAST program for nucleic acids and proteins (Table I). Six clones were identified as previously characterized genes. These known clones coded for an inwardly rectifying potassium channel (IRK-2), nucleosome assembly protein-1 (NAP-1), ribosomal protein L7, elongation factor 1␣ (EF-1␣), non-muscle myosin light chain (MLC), and heat-shock protein-90 ␤ (hsp90). Four other clones (4, 5, 137, and 169) showed significant similarities to DNA sequences in the data bases but corresponded to proteins whose functions are unknown (accession numbers: 4, Ͼgb/D14812/HUMORF16; 5, Ͼemb/Z45305/HSC2MB121; 137, Ͼemb/Z43809/HSC1LE021; 169, Ͼgb/M85812/M85812). The remaining 9 clones did not match any known gene in the data bases.
In Vivo mRNA Regulation-Total RNA was extracted from the thyroids of several control or MM/PTU-treated dogs and subjected to Northern blot analysis. Acridine orange staining, rather than ␤-actin, glyceraldehyde-3-phosphate dehydrogenase, or cyclophilin hybridization, was used to normalize the amount of total RNA transferred onto the filter. Indeed, these three genes are modulated in our system (data not shown). One representative experiment is illustrated in Fig. 1. Clone 3 mRNA is up-regulated in the stimulated dog thyroids. Clone 165 mRNA level was highly variable between dogs, enhanced in some animals or attenuated in others. It was therefore not further characterized. Clones 5, 4, 2, 45, 143 (hsp90), and 134 (NAP) transcript levels were all reduced in stimulated thyroids. No modulation was observed for clone 44 (EF-1␣) and clone 169, and no signal was detected for clones 101, 51, and 53 (IRK-2). Hybridization with clones 16, 59, 137, and 166 revealed only a smear, but no discrete transcript. All these data were confirmed in several dogs. Thus, these Northern studies demonstrate the modulation of five unknown genes in stimulated thyroids.
Analysis of Tissue Distribution-The following dog tissues were harvested shortly after sacrifice, snap-frozen in liquid nitrogen, and investigated by Northern blotting: thyroid, spleen, liver, lung, heart, skeletal muscle, kidney, adrenal, ovary, testis, stomach, pancreas, large intestine, lymph node, cerebrum, and cerebellum. Clones 2, 3, 4 (Fig. 2), 59, 165, and 169 (Table II) showed no particular tissue distribution pattern. They were detected in most of the tissues analyzed, with a particularly important expression in the adrenal for clone 3, and in the testis, thyroid, and brain for clone 4 (Fig. 2). Clone 5 showed strong thyroid specificity although it could also be clearly seen in both the ovary and cerebrum (Fig. 2). Clones 45 and 51 showed a much higher expression in the thyroid than in the other tissues. Clone 53 (IRK-2) was present only in skeletal muscle, cerebrum, and cerebellum, which was not unexpected (Table II). Clone 101, interestingly, appeared to be confined solely to the liver. To obtain qualitative estimates of the proliferative state of the tissues, those were stained with a mono- clonal antibody (PC10) against proliferating cell nuclear antigen (PCNA). The number of PCNA positive nuclei in a tissue is supposed to be closely correlated to that tissue's growth fraction (21). However, the proliferative activity, qualitatively assessed by two observers, suggests no correlation between the tissue mRNA expression level and the level of proliferation (Fig. 2, Table II).
In Vitro mRNA Regulation-In primary cultures of dog thyroid cells, the thyrocytes were cultured for 4 days without mitogenic agents to allow cells to spread and reach a quiescent state. In the absence of stimulating agents, virtually no labeling of nuclei was observed during the 48 h after the addition of [ 3 H]thymidine to the culture medium. Addition of TSH (1 milliunit/ml), forskolin (10 Ϫ5 M), an adenylate cyclase activator, EGF (25 ng/ml), and FCS (10%) at day 4 induced, after a lag of about 18 h, a progressive increase in the fraction of [ 3 H]thymidine incorporating nuclei (30). The results of Northern blot hybridization analyses are presented Fig. 3, 4, 5, 6, and in Table III.
The clone 5 mRNA levels peaked after 20 h incubation in the presence of TSH and decreased after 48 h (Fig. 3). EGF also increased clone 5 mRNA levels after 6 -24 h (Table III). Chronic stimulation by EGF and FCS did not change the expression of this clone (Fig. 3).
Clone 3 mRNA was down-regulated by forskolin (Fig. 4A) and TSH (not shown) for the first 24 h. It was increased thereafter, but never exceeded the control levels, even after a 7 days stimulation (not shown). EGF led to an up-regulation, with acute ( Fig. 4A) as well as with chronic stimulation (Table III). The mRNA content was already enhanced after 8 h of EGF stimulation, and increased further with longer incubation times. The negative effect of forskolin or TSH on mRNA levels was also observed on the EGF stimulation, since the simultaneous addition of EGF and TSH for 48 h led to a lower mRNA level than when EGF was added alone (Fig. 4B). The effects of EGF and EGF ϩ TSH on clone 3 mRNA content were reproduced by HGF (50 ng/ml), another potent mitogenic agent for thyrocytes (31), and HGF ϩ TSH. However, the increase observed with HGF was less pronounced, and the TSH inhibition was smaller (Fig. 4B).
Clone 2 mRNA levels were slightly increased after TSH treatment whereas a 3-day EGF ϩ FCS treatment led to a clear accumulation (Fig. 5, Table III). Acute stimulation by EGF also enhanced clone 2 mRNA levels ( Table III).
The mRNA levels of two identified clones (hsp90 and NAP) were also studied (Fig. 6). Treatment with forskolin led first to a decrease of hsp90 and NAP-1 mRNA levels, then after 20 h to an increase. Hsp90 mRNA control levels increased in parallel with the age of the culture; they were further enhanced by chronic treatment with EGF (Table III).
Analysis of the data obtained for the other clones presented in Table III shows that clone 4 mRNA expression was enhanced by TSH and EGF, and clone 45 mRNA amounts were increased by TSH, but remained equal after EGF treatment. As in vivo, no modulation was observed for EF-1␣ and clone 169, no signal was detected for clones 101, 51, and IRK-2, and hybridization with clones 16, 59, 137, and 166 was uninterpretable.
Characterization of Clones 5 and 3-Four clones were selected for further characterization: clones 5 and 3 (based on their modulated mRNA levels), clones 45 and 51 (based on their tissue distribution). The complete sequences of clones 5 and 3 were determined. Hydropathicity profile of the predicted 343amino acid sequence of clone 5 according to the analysis of Kyte and Doolittle (48) revealed the presence of a stretch of 20 hydrophobic amino acids, which might represent a potential transmembrane region. Protein sequence comparison of clone 5 with the data bases showed Ϸ30% identity with the protein kinase family (serine-threonine kinases). However, among the 12 well conserved subdomains of the kinase catalytic domain (32), only two are present in our clone: subdomains VIII and IX, with an homology of Ϸ80% (Fig. 7a). Clone 3 encodes a 659amino acid protein containing 6 ankyrin-like repeats (Fig. 7b). The DNA library accession numbers of clones 3 and 5 are, X99145 and X99144, respectively.

DISCUSSION
To identify new genes modulated by the TSH-dependent mitogenic pathway, we performed a differential screening of a thyroid cDNA library from a dog treated in vivo with MM/PTU. The use of an in vivo chronically stimulated thyroid insured, with no a priori assumption, that the real in vivo growth pathway was investigated. This led to the identification of 11 clones showing overexpression and 8 showing underexpression. While 6 clones encode known proteins, the 13 others have not been described to date. Among them, 4 (5, 4, 137, and 169) showed significant similarities to DNA sequences in the data bases but corresponded to proteins whose function is unknown. In vivo mRNA regulation was assessed by Northern blotting to confirm the results of the differential screening. Among the unknown candidates, clone 3 mRNA was up-regulated and clones 5, 4, 2, and 45 mRNAs were down-regulated after antithyroid drug administration. mRNA tissue distribution analysis showed neither a particular tissue distribution for most of the clones studied, nor a parallelism between mRNA expression and proliferation rate. Clones 5 and 45 are more abundantly expressed in the thyroid. Clone 51 appeared to be exclusively present in this tissue, but the signal was very faint.
Northern blotting showed that 5 unidentified cDNAs (2, 3, 4, 5, and 45) out of 13, corresponded to mRNA of which levels  were modulated by the cAMP-dependent and -independent mitogenic pathways in dog thyrocytes in primary cultures. This could result from modified transcription, degradation, or both. Remarkably, clone 3 mRNA levels were increased in vivo but not in vitro by TSH. This discrepancy suggests that the in vivo situation does not necessarily match the in vitro stimulation. Whereas in vitro the cAMP-dependent pathway alone is directly activated by TSH or forskolin, the situation in vivo appears to be more complex. The presence of EGF and IGF-1 in the thyroid has indeed been described (14), and the TSH stimulation induced in vivo by the treatment with the antithyroid drugs could indirectly involve these growth factors or even other, undescribed factors. In other words, the effects observed in vivo could be the result of different interacting signaling components, which is not the case in vitro. To confirm clone 3 mRNA up-regulation by agents acting via tyrosine kinase receptors, we investigated the action of HGF, which is the most potent mitogen for our cells (31). HGF reproduced the effects of EGF, but the increase was less pronounced. As HGF has a weaker dedifferentiating action on thyrocytes than EGF, these results may suggest that clone 3 mRNA expression is inversely related to the differentiation state of the thyrocytes. Among the 19 clones which were apparently differentially expressed in chronically stimulated thyroids, 7 (2, 3, 4, 5, 45, hsp90, and NAP-1) were regulated in vivo, as suggested by the screening, and in vitro by the various mitogenic cascades; three (101, 51, and IRK-2) were not detected by Northern blot analysis, four (16,59,137, and 166) led to uninterpretable results, and two (EF-1␣, and 169) were not regulated.
In our thyrocytes, hsp90 mRNA expression was modulated by chronic treatments: positively by EGF, negatively by TSH and forskolin. As both these agents are mitogenic, it is difficult to relate this expression to proliferation as in some other systems (33)(34)(35)(36)(37).
NAP-1 assembles the nucleosomes by mediating the formation of a histone octamer and transferring it to DNA (38,39). In our cells, NAP-1 mRNA varied in a cyclical manner in response to TSH. This could correspond to the modulation of the histones, which are well known cell cycle-dependent genes. During the cell cycle of Balb/c 3T3 cells, the content of NAP-1 was found to change in parallel with histone synthesis (40). The modulation of its expression in our cells is compatible with this concept. The expression of a human NAP-related gene also increases rapidly in T cells after induction of cell proliferation, and mitogenesis was inhibited in cells treated with the corresponding antisense oligonucleotide (41).
On the basis of the in vitro mRNA regulation results, we selected first clones 5 and 3 for further characterization. Both were completely sequenced. Clone 5 protein sequence comparison with the data bases revealed Ϸ30% identity with the protein kinase family. This large superfamily of homologous proteins contains a 250 -300-amino acid kinase domain, composed of 12 conserved subdomains that fold into a common catalytic core structure (32). However, our clone lacks most of these consensus motifs: indeed, only subdomains VIII and IX are present (with an identity of Ϸ80%). Subdomain VIII appears to play a major role in the recognition of peptide substrates, and contains several autophosphorylation sites, required for maximal kinase activity in many protein kinases. No clear ATP-binding site is identified as in the Bcr protein serine/threonine kinase (42) and in the human A6 tyrosine kinase (43). Clone 5 could represent a new example of these atypical kinases, and work is currently in progress to test this hypothesis.
Clone 3 protein sequence comparison with the data bases showed that this protein contains 6 ankyrin-like repeats. The ankyrin-like repeat is a 31-33-amino acid motif which consists of a TPLHLA core sequence and 8 -10 other well conserved residues. These motifs are found in ankyrin and several other proteins, very heterogeneous with respect to their origin, function, intracellular localization, and number of repeats. These proteins include cell cycle control proteins (e.g. cdc10 gene product of Schizosaccharomyces pombe), putative integral membrane proteins (e.g. Notch gene product of Drosophila melanogaster, involved in cellular differentiation), transcription factors (e.g. NF-B p105 precursor), and others (e.g. IBlike molecules, virally-encoded proteins) (44). More recently, ankyrin repeats have been discovered in INK4 proteins, a family of cell cycle regulators characterized by their selective inhibition of cdk4 and cdk6 (45)(46)(47). Ankyrin-like repeats are involved both in interactions between distinct proteins and within a single molecule, and this can serve a variety of functions. The presence of 6 of such motifs in clone 3 protein suggests that it might interact with other proteins, and these putative partners will be sought. We are now studying the possible role of clone 3 and clone 5 proteins in the proliferation and differentiation of the thyroid cell. The entire proteins are shown schematically. a, clone 5. Homologous regions with subdomains VIII and IX of different protein kinases are shown (accession numbers: 1, pirPA44493; 2, pirPJU0270; 3, spPP22987; 4, spPQ05512; 5, spPP18653). The consensus line is given according to the following code: uppercase letters, invariant residues; lowercase letters, nearly invariant residues; o, positions conserving nonpolar residues; ϩ, positions conserving small residues with near neutral polarity (from 32). b, clone 3. The position and the alignment of the six 31-amino acid, ankyrin-like repeats are indicated. The consensus sequence is shown below.