The sequence of a human P-tubulin cDNA clone (D,B-1) is described; our data revealed 95.6% homology compared with the sequence of a human f-tubulin processed pseudogene derived by reverse transcription of a processed mRNA (Wilde et al., Nature [London] 297:83-84, 1982). However, the amino acid sequence encoded by this cDNA showed less homology with pig and chicken 1-tubulin sequences than the latter did to each other, with major divergence within the 15 carboxy-terminal amino acids. On the other hand, an independently isolated, functionally expressed genomic human 1-tubulin sequence (5,B) possessed a very high degree of homology with chicken and pig 3-tubulins in this region. Thus, human cells appear to contain two distinct ,B-tubulin isotypes. Both the intact ,B-tubulin cDNA clone and a subclone containing only the 3' untranslated region detected two mRNA species in HeLa cells; these mRNAs were 1.8 and 2.6 kilobases long and were present in about equal amounts. Two independently subcloned probes constructed from the 3' untranslated region of the 53 genomic sequence also detected a 2.6-kilobase 1-tubulin mRNA. However, the 3'-untranslated-region probes from the cDNA clone and the genomic sequence did not cross-hybridize. Thus, at least two human ,B-tubulin genes, each specifying a distinct isotype, are expressed in HeLa cells, and the 2.6-kilobase mRNA band is a composite of at least two comigrating f-tubulin mRNAs.a-and P-tubulins are the major soluble proteins of microtubules. These long filamentous structures are found in all eucaryotic cells and are involved in a remarkable diversity of cellular functions (11). In humans, each of the genes encoding a-and ,-tubulins constitutes a large multigene family of about 15 to 20 members (8), the majority of which have been isolated in recombinant fragments (9, 41). Sequence analyses of several of these genes have shown that many of them are pseudogenes (that is, sequences which contain one or more genetic lesions that preclude the synthesis of a functional protein product). Among these pseudogenes is a novel class that is characterized by (i) the absence of intervening sequences, (ii) the presence of a short polyadenylic acid [poly(A)] tract 3' to the poly(A) signal, and (iii) the existence of short flanking direct repeat sequences (22,42,43). These pseudogenes most probably arose by reintegration into the host germ line of a cDNA transcript synthesized on a processed mRNA template.Because a significant number, perhaps the majority, of human a-and P-tubulin genes are pseudogenes, an important question concerns the number offunctionally expressed sequences.One approach to this problem is to investigate the complexity of tubulin mRNAs by constructing cDNA clones. In this paper we describe the sequence of a human P-tubulin cDNA clone that encompasses 98.4% of the coding region. Our data revealed very extensive homology with the sequence of a processed human P-tubulin pseudogene and enabled us to estimate the time of insertion of the mRNA-derived molecule into the germ li...