as a HMG box (2). Four murine autosomal SRY-related genes have been described having high homology to the HMG box region of SRY (3). We have cloned additional members of this gene family from phylogenetically diverse organisms. cDNAs prepared from mouse or human RNAs or phage lysates from cDNA libraries (Xenopus oocyte (4) and Drosophila embryo (0-12 hr) (Stratagene)), were used as templates in the polymerase chain reaction (PCR), using degenerate oligonucleotide primers which corresponded to regions conserved in the known SRYand SRY-related amino-acid sequences (Figure 1). Sequencing of the cloned PCR products demonstrated that they were heterogeneous, with the clones falling into a few, distinct classes, mostly differing from those SRY-related clones previously described (3). All of the encoded proteins are more closely related to SRYthan to other HMG-box proteins, including the T-cell factor TCF-1 (5) and the product of the S.pombe mating type gene, Mc (6). The SRYlike genes have been named SOX ('SRY-box') genes (R.LovellBadge, personal communication). Some of our mammalian SOX cDNAs encode proteins which are very similar (> 90% identity) to one or more of the Xenopus clones. It is difficult, therefore, to determine which may be orthologues. One of the Xenopus cDNAs, however, differs from the mouse and human SOX-5 sequence by only one amino acid and hence we have named it XSox-5. We may have also cloned the human orthologue of mouse Sox-4 (3), as one of our cDNAs encodes a protein which is 96% identical with the mouse sequence. SOX-5 and -6 were both cloned from mouse and human cDNA. Expression of SRY-like genes during embryogenesis in mouse (3) and Drosophila and the strong evolutionary conservation of this gene family suggests that the SRY-related genes may be of importance in developmental processes. The SOX genes are highly related in their DNA binding domain, which reflects an overlap in their sequence specificity (Ref. 7 and P.D., S.S., Frances Connor and A.A., submitted). It will be important to determine the structure of these genes outside the HMG-box, as these regions may specify other functions such as protein -protein interactions. ACKNOWLEDGEMENTS