The Q cells are migratory neuroblasts located initially between V4 and V5. They are fascinating cells because they migrate in opposite directions: QR (on the right) and its descendants migrate anteriorly whereas QL and one of its descendants migrate posteriorly (see Figure). Thus, by studying mutations that affect Q migration, it may be possible to identify genes involved in distinguishing left from right and anterior from posterior as well as genes involved in the process of cell migration itself. We have begun to analyze two genes that affect Q cell migration,
mab-5 and
mig-1.
mab-5 encodes a homeodomain protein required for many posterior cell types to develop correctly. In particular, it is required for QL to migrate posteriorly. Mosaic analysis suggests that
mab-5 is required in a cell-autonomous fashion for Q cell migration (see Kenyon, Cell 46, 477), and not in the cells that contact QL as it migrates (with the possible exception of V5L, sister of QL). M. Chalfie (1983) found that in
mab-5(
e1239) mutants, QL migrates posteriorly as in wild type but its two daughter cells reverse direction and migrate anteriorly. We have followed Q migration in animals carrying a putative gain-of-function allele of
mab-5 originally isolated and characterized by E. Hedgecock (molecular analysis of this allele will be reported in a future WBG article by Salser and Kenyon). In this
mab-5(gf) mutant,
e1751, descendants of both QL and QR (AVM/PVM and SDQL/R neurons) are located in the posterior (Hedgecock, pers. comm.). We followed the migration of QR in one
mab-5(gf); 90) animal and observed that, again, the initial phase of migration was normal (QR migrated anteriorly) but that the forward migration stopped after the first cell division. After additional cell division, QR.ap migrated back toward the tail. From these observations, it seems that control of QL migration is complicated and has several distinct phases. The initial phase of QL migration, prior to division, does not require
mab-5; therefore, at least initially,
mab-5 is not required for distinguishing left from right. Instead,
mab-5 seems to have two functions in Q migration. First, it appears to prevent the daughters of QL from migrating anteriorly. Later,
mab-5 seems to cause QL.ap to migrate posteriorly into the tail region instead of anteriorly to the head. The
mig-1 gene, which is also required for correct Q cell migration, was identified independently by Hedgecock (1987) and Desai (1988). In
mig-1(
e1787) mutants, descendants of both QL and QR (AVM/PVM and SDQR/L) are located in the anterior body region (10/20 SDQL and PVM cells). We followed QL migration in
mig-1(
e1787) and found that the QL migration pattern resembles that of loss-of-function
mab-5 mutants ( 2/2); that is, QL migrates posteriorly but its daughters reverse direction and migrate anteriorly. Therefore, it seems probable that
mab-5 and
mig-1 are required for the same phases of the migration of QL descendants. We were interested in learning whether the Mig-1 phenotype could be suppressed by the
mab-5(gf) mutation. Therefore,
mig-1(
e1787);
mab-5(
e1751) double mutants were constructed. In these animals,
mab-5(gf) suppresses
mig-1 completely: QL descendants PVM and SDQL retain their wild type posterior positions (20/20), and QR descendant AVM and SDQR are also located posteriorly (21/22). The QR migration pattern was identical in
mab-5(gf) and
mig-1,
mab-5(gf) animals (2/2). The most straightforward interpretation of this result is that the
mig-1 product acts upstream of
mab-5 to activate
mab-5 in the daughters of QL. However,
mig-1 could act downstream of
mab-5 if mig- 1
(e1787) is not null, and the
mab-5(gf) mutation enhances residual mig- 1 activity. Alternatively,
mig-1 and
mab-5 could act independently to promote posterior migration, and the
mab-5(gf) product could allow the cell to bypass the requirement for the
mig-1 product. Information about the
mig-1 null phenotype and
mab-5 expression patterns in
mig-1 mutants will help distinguish between these possibilities. [See Figure 1]