(i) Phylogenetic analysis:
     Extensive BLAST searching of the A. thaliana genome yielded 18 putative ß-galactosidase genes; these were designated AtBGAL1 through AtBGAL18. AtBGAL18 appears to be a pseudogene. Sequence analysis of the remaining 17 family members was performed at UI. Several major features of the putative A. thaliana ß-galactosidase proteins were identified by analyzing their deduced polypeptide sequences. First, all Arabidopsis ß-galactosidases except AtBGAL7 and AtBGAL16 possess N-terminal signal peptides. Second, all Arabidopsis ß-galactosidase protein sequences include the putative active-site consensus motif G-G-P-[LIVM]₂-X₂-Q-X-E-N-E-[FY], within which the second glutamate residue is believed to be a catalytic residue. Third, ten of the ß-galactosidases (AtBGALs 1, 3, 7-9, 11, 13-16) contain C-terminal lectin-like domains. It is unknown whether these domains are functional, but their existence and conserved nature raise the interesting possibility that they may increase the catalytic efficiency of ß-galactosidases by anchoring these lectin-possessing hydrolases to their polymeric substrates. Finally, as predicted by PSort, 12 Arabidopsis ß-galactosidases may be targeted to the cell exterior, suggesting their potential involvement in cell wall modification. These hydrolases are probably glycoproteins, given their possession of multiple (n =1-9) potential N-glycosylation sites.

     Phylogenetic analysis of the Arabidopsis ß-galactosidase gene family was performed by first using ClustalX to align the polypeptide sequences and then using PAUP* 4.0 to construct a Parsimony tree. The robustness of the nodes within this tree was tested by a bootstrap resampling method. The resulting tree indicated that, with the exception of AtBGAL17, all Arabidopsis BGAL genes formed a monophyletic group that is separated from BGAL genes from bacteria, fungi, and animals. Five major clades were formed among the 16 AtBGAL genes in the Arabidopsis group. The same five clades were supported in a second tree constructed with AtBGAL17 as an outgroup. Based on these data, AtBGALs 1-5, 7, 12, and 15, which constitute two major clades, were assigned to VPI, with the remainder assigned to UI.

(ii) Obtaining cDNAs Encompassing the ORFs of Mature Proteins:
     To facilitate cloning of the nine AtBGAL genes assigned to UI, we decided to determine the expression patterns of these genes, before attempting to clone each gene from cDNA populations prepared from the specific organ in which highest expression levels were observed. Total RNA preparations were isolated from six Arabidopsis organs (etiolated seedlings, light-grown seedlings, leaves, roots, flowers, and siliques). Relative RT-PCR analyses were performed using the ATP synthase ß-subunit gene as the internal standard. Gene-specific fragments were amplified from all six sources and resolved by gel electrophoresis. The results indicated that all genes except AtBGAL9 yielded strong PCR products with flower as the source of mRNA (data not shown). In addition, some genes (AtBGAL8, AtBGAL10, and AtBGAL17) are expressed in at least four different organs, whereas expression of AtBGAL11, and perhaps also of AtBGAL16 and AtBGAL13, was detected only in flowers.

     Based on these expression patterns, we went on to prepare full-length cDNA populations from flower, leaf, and root poly(A)⁺ RNA according to a modification of the SMART-cDNA synthesis procedure (Clontech). Gene-specific primers were used in PCR to obtain cDNA clones that encompass the complete ORF of mature peptides for each AtBGAL gene. We succeeded in obtaining cDNA clones for AtBGALs 6, 8, 9, 10, 11, 13, 16, and 17. The identity of each clone was confirmed by DNA sequencing. In addition, our results indicated that the exon-intron assignments for AtBGAL6 and AtBGAL17 posted in Genbank were erroneous. The sequences of our AtBGAL cDNA clones will be submitted to Genbank soon.

     A summary of our progress toward isolating target ß-galactosidase cDNAs and expressing the encoded ß-galactosidases in P. pastoris is provided in Table II. We have been unable to obtain the desired AtBGAL14 cDNA from either flower or root cDNA libraries and are currently screening other cDNA libraries for this gene.